An AI workspace that operates on the business, not on screenshots of it
Drop in a competitor ad. The Space turns it into objects, links them with typed edges, lets two people and an agent work the same ids, projects its numbers, opens issues on them and keeps watching on a rule. Every card on this page is the live object, in the look of its app.
Drop in a competitor ad. The Space turns it into objects, links them, and keeps watching.
The overview is the graph again, as numbered panels: the input, the analysis, the competitor, its pages, the variants, the graph, a direct reply field, the differences beside us (unknowns are said, never filled), what is not known (each a need a person or an agent can take), the rules as switches, the metrics, the timeline, the apps the engine registers, the people and the agent. Every panel reads the same ids.
The address https://relevanceai.com/ was dropped into the Space on 2026-10-09. The build fetched the page itself (sha256 ade5239277240a02…, 301339 bytes) and made these objects, each with its own id and its own page:
assetobj_6d0c383a011ce1d35583e74b: the original, with its hash and the exact fetch
adobj_6df3cb08754fb4008f21b3be: what the page presents as its ad (headline, text, image, call to action); labelled as such
competitorobj_9a90225482d0d81ed742c2c1: Relevance AI, resolved from the landing page’s domain, never from a name alone; the evidence is on the edge
pages ×6 and research ×1: the competitor’s public pages (each with its hash) and public search results, labelled as search results
analysisobj_a0e0c09160cba70ca3d08e8a: nine statements, each extracted with its source, confirmed, rejected or left open
claims ×2: sentences the page makes, each an object the ad contains_claim
hypotheses ×3 and variants ×3: each variant derived_from the ad and tests one hypothesis
ruleobj_7056be75b84c01cb84dcb14e: a sentence turned into a structure the runtime enforces
Every line on the board is a typed edge with its evidence, author, time and version. Click one.
The ad, as an object
Correct one statement
The analysis is not a paragraph. It is nine statements, each marked extracted (with the words it was read from), confirmed, rejected or open. Anyone reading this page may confirm, correct or reject a statement here; the change is recorded with who made it, the earlier value is kept, and the same statement changes on the board above and on the object’s own page.
The same object from a reply
A reply addresses the object by its id, not a recap of a chat. The Space’s agent reads the words: find their other public pages, make three variants, check this competitor daily are operations; anything else is kept as a note on the object. What the reply made appears in the Space, not only here.
The sentence “Check this competitor daily. When a new offer appears, compare it with ours and propose three alternatives. Do not publish anything without approval. Only generate alternatives if they sell a product we also sell.” became version 1 of a rule: every day at 09:00; if offer.changed eq true and product.overlap eq true; then observe, compare_offer, propose_variants ×3; gates: approval before publish; approval before send; budget: max_fetches 4, model_calls 0, max_variants 3, max_runs_per_day 1. A dry run reads what the Space knows and spends nothing. A new sentence makes a new version with its difference against the one before. Activation and checks are gated by role; this build never publishes or sends from a rule.
A metric is a number with a source and a date; this page invents none. Below, two kinds are real to this build: the views of six articles of this site (its own counter), and a base the owner typed in as an example, labelled as such. A projection is an object: the sentence “if CTR goes up 10% and CPC down 5%” became assumptions; the funnel was recomputed with the formulas written in the object (clicks = impressions × CTR, spend = clicks × CPC, conversions = clicks × CVR, revenue = conversions × AOV); every row names its measured source, its formula or its assumption; rows that need a metric nobody measured stay empty and say so. Toggle Traffic, Spend, Conversion or Market; transpose rows and columns; open the JSON under the view; reply to the projection to change an assumption (a new version, with its history).
A view is a saved definition over the same objects: what to show, grouped by which field, filtered how, sorted how, drawn as a table, a board, a chart, a funnel or the apps. It is made by a sentence through one grammar (show | table | board | chart | graph | list <what> [by <field>] [where <field> <op> <value>] [sorted by <field>] [as funnel|bars|lines] · compare <a> and <b> · levers <kinds> · arrange | group | filter | sort … · apps · what is missing to compare <metric> · if <scenario>, show the funnel). What the grammar does not understand is said back, never guessed: “show the flibbertigibbet by moon phase” answers “not understood: by moon phase / flibbertigibbet”. A model saying any of this says it through the same grammar; no model was called to make these.
The sentences below were typed into this Space on 2026-10-09; each saved view, its id and what the grammar read are in facts.views_from_sentences. The levers on the Metrics view above are sliders: each writes a new projection version when it settles, and two moved together write one version; a projected row opens into the intermediate variables and the formula that produced it, each input marked measured, assumed, derived or missing.
“apps” — every object of the Space that lives in an app, drawn as that app: the issue as a Reminders card, the pages as Safari, the screenshots as Photos.
An issue is a task of this build (its Tasks tab), opened from the Space with a zero-context spec composed from the object it is about, attached to the Space and linked to that object with an about edge. The board of issues is the Tasks list by state.
Every ingest, resolution, interpretation, correction, proposal, dry run, activation and refusal is an event with its actor, the object’s version and its Ledger row.
the proof: two signed-in people, one object, a change seen by both, a refusal
JSONthe proof: two signed-in people, one object, a change seen by both, a refusal
{9 keys
"what_this_is": "two separately authenticated humans and the deterministic Space agent on one canonical object graph; every step below is a live call the build answered, with its receipt",
"human_a": {3 keys
"who": "the owner",
"how": "the build key (x-sheets-key), the same credential the Console uses",
"viewer": {2 keys
"kind": "owner",
"id": "owner"
}
},"human_b": {6 keys
"who": "Space tester (member)",
"person_id": "p_n5cauqlp",
"how": "a person record added by the owner (POST /api/people), a one-time invitation link made with link_only (not sent by the build; the owner hands it over), signed in by POST to that link: an ms_session cookie of their own",
"role_in_space": "editor",
"ops_granted": [6 items
"read",
"analysis.confirm",
"variant.review",
"policy.revise",
"policy.dryrun",
"space.layout"
],"viewer": {2 keys
"kind": "person",
"id": "p_n5cauqlp"
}
},"agent": {4 keys
"id": "space-agent",
"kind": "deterministic",
"what_it_did": [3 items
"resolved the advertiser from the landing page’s domain (ad.resolve, ex_b88f6460728fe2b33e22f4ce9db2)",
"wrote the analysis (ad.interpret, ex_f0a63b467224c58e054129995ba6)",
"read the watched pages twice on the owner’s press and recorded two observations (policy.check, ex_ab497a2a77c3836604e285791bef and ex_553e759661f9df0a7058c3663d5f)"
"did": "POST /live/objects/obj_a0e0c09160cba70ca3d08e8a/ops/analysis.confirm {field: offer, decision: reject, note: the figures $7, $48, $1.2 are per-task costs in the page text, not the offer the ad makes}",
"did": "policy.activate (ex_dcebeaccfb4860c8112159fbfc18): active, next check 2026-10-09T09:00:00Z; then policy.check twice on the owner’s press",
"observations": [2 items
"obj_0d4432a7d566c0872f27e2ec: nothing changed on 4 watched pages (first reading)",
"obj_483635a9343e0b3becfc4671: nothing changed on 4 watched pages"
],"dry_run_after": {3 keys
"facts": {4 keys
"ad.count": 1,
"page.changed": false,
"offer.changed": false,
"product.overlap": false
},
"holds": false,
"would_run": []
}
},{4 keys
"step": 10,
"who": "a visitor (no sign-in)",
"before_sharing": {2 keys
"GET /live/s/….json": 403,
"POST analysis.confirm": 403
},"after_the_owner_shared": {3 keys
"GET /live/s/….json": {2 keys
"me": {2 keys
"kind": "guest",
"role": "viewer"
},
"objects": 22
},
"POST analysis.confirm (field cta, confirm)": "succeeded, ex_9c2394eaa1d2218376b771751def: the owner granted * [read, analysis.confirm] on this one object",
"where_the_decision_is_made": "src/blocks.js authorize(), on the server, from the object’s grants and the viewer the request proves (key, session cookie, or a visitor cookie); the widgets only draw what the server offers",
"synchronization": "polling: every surface reads the graph again every 20 seconds (not while someone types) and at once after any widget on the page changes an object; no push",
"not_claimed": "cryptographic non-repudiation. Each event carries the actor the server resolved and a Ledger row id; the Ledger is append-only on the build, not independently witnessed."
}
What is real here, and what is not
real fetches, simulated data, model calls
JSONreal fetches, simulated data, model calls
{7 keys
"real_fetches": "every page in this Space was read by the build’s WEB_FETCH tool and kept with its sha256; the asset is the landing page as fetched",
"simulated_data": "none",
"the_ad": "the landing page’s own presentation of itself (og:title, og:description, og:image, its first call to action), labelled so on the ad object; a Meta Ad Library address is accepted but its content cannot be read from the cloud (Meta answers the cloud browser with a bot challenge; recorded on the asset when it happens)",
"model_calls_in_the_pipeline": 0,
"inferences": "none made: the two statements that need one (positioning, audience) are left open and say so; the model path is the owner’s chat, where this object travels with the words",
"variant_images": "not rendered: variant.render is wired to Workers AI and runs only on the owner’s press; the variants here are objects with a headline, a prompt, a hypothesis and a review state",
"public_search_results": "stand in for an ad library and are labelled as such on the research object"
}
where it stops
JSONwhere it stops
[7 items
"Meta Ad Library pages are not readable from the cloud (bot challenge); the Mac scraper opens the owner’s own Chrome and was not run",
"two of nine statements of the analysis are open (positioning, audience): they need a model, which is not run here",
"the rule’s clock is read as UTC; the scheduled pass runs only when the switch spaces-monitor is on (off by default)",
"edge access scope is recorded (members) but every edge of a readable space is shown; per-edge disclosure is not enforced yet",
"the example base metrics were typed by the owner as an example and are labelled so on each metric; the site’s view counts are the only measured numbers here",
"Meta Ads tools of this build do not reach a service today (META_ADS_ACCOUNTS answers “no way that reaches a service”), so no ad-account metric is imported",
"offers: the competitor’s pricing page carries no prices in its HTML, so 0 products were made; the figures the landing page carries were extracted as “offer” and a person rejected the statement"
]
The object graph behind the page
A Space is itself an object (type space, adapter spaces). Its members are grants on the space and on everything it contains; a role (viewer, editor, approver) is a list of operations. Objects live in ms.app_objects; edges in ms.app_edges (source, target, type, evidence, author, version, access); events in ms.app_events, each with a Ledger row. Every operation runs through the live-objects engine: authorized on the server for this viewer, written as an intent first, receipted. The graph is read at GET /live/s/<space id>.json by whoever opens it, under their own permissions.
the graph today
JSONthe graph today
{4 keys
"objects": 31,
"edges": 62,
"events": 35,
"by_type": {15 keys
"task": 1,
"need": 1,
"projection": 1,
"metric": 6,
"observation": 2,
"policy": 1,
"variant": 3,
"hypothesis": 3,
"research": 1,
"page": 6,
"claim": 2,
"analysis": 1,
"competitor": 1,
"ad": 1,
"asset": 1
}
}
edges by type
JSONedges by type
{14 keys
"contains": 31,
"evidenced_by": 1,
"publishes": 1,
"links_to": 1,
"owns": 6,
"interprets": 1,
"contains_claim": 2,
"about": 2,
"derived_from": 3,
"tests": 3,
"watches": 1,
"observed": 2,
"triggers": 2,
"projects": 6
}
events, by kind, with Ledger rows
JSONevents, by kind, with Ledger rows
{15 keys
"denied": [3 items
{6 keys
"id": "ev_1eb31fa9e487472063f5",
"object": "obj_7056be75b84c01cb84dcb14e",
"actor": "a visitor (guest:78f2056f10b86b25d3408189)",
{
"schema":"https://miscsubjects.com/schemas/jarticle.json",
"status":"published",
"slug":"live-business-spaces",
"title":"An AI workspace that operates on the business, not on screenshots of it",
"kind":"procedure",
"summary":"Drop in a competitor ad. The Space turns it into objects, links them with typed edges, lets two people and an agent work the same ids, projects its numbers, opens issues on them and keeps watching on a rule. Every card on this page is the live object, in the look of its app.",
"tags": [
"software-objects",
"ontology",
"competitive-intelligence",
"spaces"
],
"facts": {
"space": {
"id":"obj_c78144b0b422f79531d614fd",
"name":"Competitive intelligence",
"link":"https://miscsubjects.com/live/s/obj_c78144b0b422f79531d614fd",
"json":"https://miscsubjects.com/live/s/obj_c78144b0b422f79531d614fd.json",
"members": [
{
"identity":"owner",
"role":"owner",
"at":"2026-10-09T06:46:34.308Z"
},
{
"identity":"p:p_n5cauqlp",
"role":"editor",
"at":"2026-10-09T06:49:14.211Z"
},
{
"identity":"*",
"role":"viewer",
"at":"2026-10-09T06:55:59.755Z"
}
],
"agent": {
"id":"space-agent",
"kind":"deterministic",
"what":"On a new asset: resolve the advertiser, interpret the ad. On a policy’s clock: read the watched pages, record an observation, propose what the rules allow. Never a model call; never a send or a publish without a person’s approval.",
"watches": [
"asset",
"policy"
]
}
},
"what_was_dropped": {
"url":"https://relevanceai.com/",
"kind":"landing_page",
"sha256":"ade5239277240a029f857f5718cdecdc658c2b42fbf6acb7a4c3bdf6581a5080",
"bytes":301339,
"fetch":"real fetch (WEB_FETCH from this build)",
"read_at":"2026-10-09T06:46:51.552Z"
},
"objects": {
"observation": [
{
"id":"obj_0d4432a7d566c0872f27e2ec",
"version":"bb37a689596d",
"link":"https://miscsubjects.com/live/o/obj_0d4432a7d566c0872f27e2ec",
"title":"nothing changed on 4 watched pages"
},
{
"id":"obj_483635a9343e0b3becfc4671",
"version":"39558ba9564e",
"link":"https://miscsubjects.com/live/o/obj_483635a9343e0b3becfc4671",
"title":"nothing changed on 4 watched pages"
}
],
"policy": [
{
"id":"obj_7056be75b84c01cb84dcb14e",
"version":"4af25c77fcfe",
"link":"https://miscsubjects.com/live/o/obj_7056be75b84c01cb84dcb14e",
"title":"Watch Relevance AI"
}
],
"variant": [
{
"id":"obj_fd2e7fa3b56c7a4d2b738fe0",
"version":"3fb3ad53c787",
"link":"https://miscsubjects.com/live/o/obj_fd2e7fa3b56c7a4d2b738fe0",
"title":"Your week back: Specialist AI Agents for Every Task | Relevance AI"
},
{
"id":"obj_7c5402d45a637a74a75963d8",
"version":"bb8a95fc4c5f",
"link":"https://miscsubjects.com/live/o/obj_7c5402d45a637a74a75963d8",
"title":"Week 3+ Deploy your first team of agents Using Invent, build your…"
},
{
"id":"obj_6741ffc7a689f9d4796245d1",
"version":"1222c9515241",
"link":"https://miscsubjects.com/live/o/obj_6741ffc7a689f9d4796245d1",
"title":"$7, $48, $1.2. Specialist AI Agents for Every Task | Relevance AI"
}
],
"hypothesis": [
{
"id":"obj_2f16a1753970252235bef4e4",
"version":"a06fb685895a",
"link":"https://miscsubjects.com/live/o/obj_2f16a1753970252235bef4e4",
"title":"Outcome over feature"
},
{
"id":"obj_f67d3c4096511b9f9e2d635a",
"version":"2f1c7d9f46cb",
"link":"https://miscsubjects.com/live/o/obj_f67d3c4096511b9f9e2d635a",
"title":"One number, one claim"
},
{
"id":"obj_047271c95aab63773d60b22e",
"version":"1643715331da",
"link":"https://miscsubjects.com/live/o/obj_047271c95aab63773d60b22e",
"title":"Lead with the offer"
}
],
"research": [
{
"id":"obj_de749ad98d2a4bcacfc26f0d",
"version":"e21b0090cef1",
"link":"https://miscsubjects.com/live/o/obj_de749ad98d2a4bcacfc26f0d",
"title":"\"Relevance AI\" ads"
}
],
"page": [
{
"id":"obj_d506e03c276ca2d8b795510b",
"version":"4e92af645ec9",
"link":"https://miscsubjects.com/live/o/obj_d506e03c276ca2d8b795510b",
"title":"Rakuten Advertising - Customer Story | Relevance AI"
},
{
"id":"obj_045acfcf0efded0a54b5458b",
"version":"e6526e39190c",
"link":"https://miscsubjects.com/live/o/obj_045acfcf0efded0a54b5458b",
"title":"Lightspeed Commerce - Customer Story | Relevance AI"
},
{
"id":"obj_11b753651a13aa03f8e9ea8e",
"version":"4dfc66be14cd",
"link":"https://miscsubjects.com/live/o/obj_11b753651a13aa03f8e9ea8e",
"title":"Autodesk - Customer Story | Relevance AI"
},
{
"id":"obj_9927d7070a1a1e5f4405027c",
"version":"16eb8da188e9",
"link":"https://miscsubjects.com/live/o/obj_9927d7070a1a1e5f4405027c",
"title":"Canva - Customer Story | Relevance AI"
},
{
"id":"obj_0cae3933f3fa6b36e88c183e",
"version":"21a334baf413",
"link":"https://miscsubjects.com/live/o/obj_0cae3933f3fa6b36e88c183e",
"title":"Enterprise AI Agents | Relevance AI"
},
{
"id":"obj_ae8fbe31f01d1d2833bfab40",
"version":"e7ac3c511c3b",
"link":"https://miscsubjects.com/live/o/obj_ae8fbe31f01d1d2833bfab40",
"title":"Specialist AI Agents for Every Task | Relevance AI"
}
],
"claim": [
{
"id":"obj_9ca6c5d94e81dd0dbd900276",
"version":"b5c1fdf352dc",
"link":"https://miscsubjects.com/live/o/obj_9ca6c5d94e81dd0dbd900276",
"title":"Contract Reviewer Wk 7 Invoice Matcher Wk 8 Week 4+ Evaluate your pilot."
},
{
"id":"obj_fa5b48fe2b6508a6487f88ff",
"version":"218912045681",
"link":"https://miscsubjects.com/live/o/obj_fa5b48fe2b6508a6487f88ff",
"title":"Week 3+ Deploy your first team of agents Using Invent, build your first team of highly customized agents."
}
],
"analysis": [
{
"id":"obj_a0e0c09160cba70ca3d08e8a",
"version":"7ad8bb0474fb",
"link":"https://miscsubjects.com/live/o/obj_a0e0c09160cba70ca3d08e8a",
"title":""
}
],
"competitor": [
{
"id":"obj_9a90225482d0d81ed742c2c1",
"version":"4998509197dd",
"link":"https://miscsubjects.com/live/o/obj_9a90225482d0d81ed742c2c1",
"title":"Relevance AI"
}
],
"ad": [
{
"id":"obj_6df3cb08754fb4008f21b3be",
"version":"7d04c5d00c69",
"link":"https://miscsubjects.com/live/o/obj_6df3cb08754fb4008f21b3be",
"title":"Specialist AI Agents for Every Task | Relevance AI"
}
],
"asset": [
{
"id":"obj_6d0c383a011ce1d35583e74b",
"version":"2ff1f4aa9e3b",
"link":"https://miscsubjects.com/live/o/obj_6d0c383a011ce1d35583e74b",
"title":"Specialist AI Agents for Every Task | Relevance AI"
}
]
},
"counts": {
"objects":31,
"edges":62,
"events":35,
"by_type": {
"task":1,
"need":1,
"projection":1,
"metric":6,
"observation":2,
"policy":1,
"variant":3,
"hypothesis":3,
"research":1,
"page":6,
"claim":2,
"analysis":1,
"competitor":1,
"ad":1,
"asset":1
}
},
"edges_by_type": {
"contains":31,
"evidenced_by":1,
"publishes":1,
"links_to":1,
"owns":6,
"interprets":1,
"contains_claim":2,
"about":2,
"derived_from":3,
"tests":3,
"watches":1,
"observed":2,
"triggers":2,
"projects":6
},
"edge_fields": [
"source",
"target",
"type",
"evidence",
"author",
"timestamp",
"version",
"access"
],
"the_ad": {
"id":"obj_6df3cb08754fb4008f21b3be",
"headline":"Specialist AI Agents for Every Task | Relevance AI",
"primary_text":"Deliver maximum ROI through optimized specialist agents. Build agents for sales, customer success, marketing and HR, and run millions of daily tasks on one platform.",
"cta":"Talk to sales",
"image":"https://cdn.relevanceai.com/images/og-image.webp",
"landing_url":"https://relevanceai.com/",
"source":"landing page (real fetch): the ad as the page presents itself (og:title, og:description, og:image, its first call to action)",
"resolution": {
"at":"2026-10-09T06:46:52.735Z",
"name":"Relevance AI",
"domain":"relevanceai.com",
"status":"resolved",
"created":true,
"evidence": {
"rule":"the registrable domain of the ad’s landing page names the advertiser; a name alone would not",
"quote":"Relevance AI",
"domain":"relevanceai.com",
"from_url":"https://relevanceai.com/",
"page_title":"Specialist AI Agents for Every Task | Relevance AI",
"og_site_name":"Relevance AI"
},
"competitor":"obj_9a90225482d0d81ed742c2c1"
}
},
"the_competitor": {
"id":"obj_9a90225482d0d81ed742c2c1",
"name":"Relevance AI",
"domain":"relevanceai.com",
"evidence": {
"rule":"the registrable domain of the ad’s landing page names the advertiser; a name alone would not",
"quote":"Relevance AI",
"domain":"relevanceai.com",
"from_url":"https://relevanceai.com/",
"page_title":"Specialist AI Agents for Every Task | Relevance AI",
"og_site_name":"Relevance AI"
},
"pages":6,
"offers_found":0,
"research":"obj_de749ad98d2a4bcacfc26f0d"
},
"the_analysis": {
"id":"obj_a0e0c09160cba70ca3d08e8a",
"statements": [
{
"key":"hook",
"status":"confirmed",
"value":"Specialist AI Agents for Every Task | Relevance AI",
"evidence":"https://relevanceai.com/",
"by":"Space tester (member) (person)",
"history":1
},
{
"key":"primary_text",
"status":"extracted",
"value":"Deliver maximum ROI through optimized specialist agents. Build agents for sales, customer success, marketing and HR, and run millions of daily tasks on one platform.",
"evidence":"https://relevanceai.com/",
"by":"the Space agent (deterministic extraction)",
"history":0
},
{
"key":"offer",
"status":"rejected",
"value":"$7, $48, $1.2",
"evidence":"https://relevanceai.com/",
"by":"Space tester (member) (person)",
"history":1
},
{
"key":"cta",
"status":"confirmed",
"value":"Talk to sales",
"evidence":"https://relevanceai.com/",
"by":"a visitor (guest)",
"history":1
},
{
"key":"claims",
"status":"extracted",
"value": [
"Week 3+ Deploy your first team of agents Using Invent, build your first team of highly customized agents.",
"Contract Reviewer Wk 7 Invoice Matcher Wk 8 Week 4+ Evaluate your pilot."
],
"evidence":"https://relevanceai.com/",
"by":"the Space agent (deterministic extraction)",
"history":0
},
{
"key":"landing_consistency",
"status":"extracted",
"value":"consistent: the landing page’s title carries the headline",
"evidence":"https://relevanceai.com/",
"by":"the Space agent (deterministic extraction)",
"history":0
},
{
"key":"visual_structure",
"status":"extracted",
"value":"one image (the page’s og:image), not read for composition",
"evidence":"https://relevanceai.com/",
"by":"the Space agent (deterministic extraction)",
"history":0
},
{
"key":"positioning",
"status":"open",
"value":null,
"evidence":null,
"by":"the Space agent (deterministic extraction)",
"history":0
},
{
"key":"audience",
"status":"open",
"value":null,
"evidence":null,
"by":"the Space agent (deterministic extraction)",
"history":0
}
],
"statuses": [
"extracted",
"inferred",
"confirmed",
"rejected",
"open"
]
},
"the_rule": {
"id":"obj_7056be75b84c01cb84dcb14e",
"text":"Check this competitor daily. When a new offer appears, compare it with ours and propose three alternatives. Do not publish anything without approval. Only generate alternatives if they sell a product we also sell.",
"version":2,
"rules": {
"gates": {
"publish":null,
"approval_for": [
"publish",
"send"
],
"variants_only_if_conditions":true
},
"scope": {
"competitors":"listed"
},
"budget": {
"max_fetches":4,
"model_calls":0,
"max_variants":3,
"max_runs_per_day":1
},
"actions": [
{
"kind":"observe"
},
{
"kind":"compare_offer"
},
{
"n":3,
"kind":"propose_variants"
}
],
"trigger": {
"at":"09:00",
"kind":"daily"
},
"conditions": [
{
"op":"eq",
"field":"offer.changed",
"value":true
},
{
"op":"eq",
"field":"product.overlap",
"value":true
}
]
},
"unparsed": [],
"summary":"every day at 09:00; if offer.changed eq true and product.overlap eq true; then observe, compare_offer, propose_variants ×3; gates: approval before publish; approval before send; budget: max_fetches 4, model_calls 0, max_variants 3, max_runs_per_day 1",
"active":true,
"revisions": [
{
"version":1,
"by":"the owner",
"at":"2026-10-09T06:48:04.944Z",
"diff":null
},
{
"version":2,
"by":"Space tester (member) (person)",
"at":"2026-10-09T06:55:11.972Z",
"diff": {
"added": [
{
"condition": {
"op":"eq",
"field":"product.overlap",
"value":true
}
}
],
"changed": [
{
"to":true,
"from":null,
"gate":"variants_only_if_conditions"
}
],
"removed": []
}
}
],
"last_dry_run": {
"at":"2026-10-09T06:55:57.652Z",
"by":"the owner",
"note":"a dry run reads what the Space already knows; it fetches nothing, spends nothing, publishes nothing and sends nothing",
"sent":0,
"spent":0,
"fetched":0,
"published":0,
"per_competitor": [
{
"name":"Relevance AI",
"facts": {
"ad.count":1,
"page.changed":false,
"offer.changed":false,
"product.overlap":false
},
"competitor":"obj_9a90225482d0d81ed742c2c1",
"evaluation": {
"holds":false,
"unknown":false,
"would_run": [],
"conditions": [
{
"op":"eq",
"field":"offer.changed",
"value":true,
"result":"fail",
"value_seen":false
},
{
"op":"eq",
"field":"product.overlap",
"value":true,
"result":"fail",
"value_seen":false
}
],
"blocked_by_gates": []
},
"pages_known":6,
"last_observation":"obj_0d4432a7d566c0872f27e2ec"
}
]
},
"runs": [
{
"at":"2026-10-09T06:55:48.219Z",
"by":"the owner",
"fetched":4,
"version":2,
"proposed":0,
"observations":true
},
{
"at":"2026-10-09T06:55:52.572Z",
"by":"the owner",
"fetched":4,
"version":2,
"proposed":0,
"observations":true
}
]
},
"events": {
"denied": [
{
"id":"ev_1eb31fa9e487472063f5",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"a visitor (guest:78f2056f10b86b25d3408189)",
"at":"2026-10-09T06:56:02.386Z",
"ledger":"worker:misc-sheets:be93ec73-ddde-438e-ae45-f6ec8c96fa71:1",
"version":"4af25c77fcfe"
},
{
"id":"ev_ad03d633059cbe695f60",
"object":"obj_9a90225482d0d81ed742c2c1",
"actor":"Space tester (member) (person:p_n5cauqlp)",
"at":"2026-10-09T06:55:12.631Z",
"ledger":"worker:misc-sheets:5267144c-1c0f-40ee-a7e4-353e6291a656:1",
"version":"4998509197dd"
},
{
"id":"ev_bb55d4fe8c92b21200e6",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"Space tester (member) (person:p_n5cauqlp)",
"at":"2026-10-09T06:55:11.452Z",
"ledger":"worker:misc-sheets:479bf1e6-a483-40cc-8113-53f34f0c8683:1",
"version":"d706bfb4560e"
}
],
"corrected": [
{
"id":"ev_d45666a06bc00b5eefb1",
"object":"obj_a0e0c09160cba70ca3d08e8a",
"actor":"a visitor (guest:78f2056f10b86b25d3408189)",
"at":"2026-10-09T06:56:01.775Z",
"ledger":"worker:misc-sheets:1ab6ffc0-1377-4642-88af-4edbd4e42417:2",
"version":"405141e27e3c"
},
{
"id":"ev_e1588206d72560bdba84",
"object":"obj_a0e0c09160cba70ca3d08e8a",
"actor":"Space tester (member) (person:p_n5cauqlp)",
"at":"2026-10-09T06:55:10.770Z",
"ledger":"worker:misc-sheets:4b398714-831e-4a6e-8438-0ea9aaf5233e:2",
"version":"7e838d47712c"
},
{
"id":"ev_6fcb190f77cb134161a0",
"object":"obj_a0e0c09160cba70ca3d08e8a",
"actor":"Space tester (member) (person:p_n5cauqlp)",
"at":"2026-10-09T06:54:31.072Z",
"ledger":"worker:misc-sheets:3898e13f-f8dd-4b4f-8825-c21f5b5cfacf:2",
"version":"1e6cf4747b06"
}
],
"shared": [
{
"id":"ev_3fac1e0aa1d697ea5655",
"object":"obj_a0e0c09160cba70ca3d08e8a",
"actor":"the owner",
"at":"2026-10-09T06:56:00.498Z",
"ledger":"worker:misc-sheets:56526fbd-00b4-4594-9fbe-e959554e3c9b:1",
"version":"405141e27e3c"
}
],
"member_added": [
{
"id":"ev_511abb00af449a905b22",
"object":"obj_c78144b0b422f79531d614fd",
"actor":"the owner",
"at":"2026-10-09T06:55:59.830Z",
"ledger":"worker:misc-sheets:771970eb-4a96-4314-89b0-ad35b973a577:2",
"version":"70a7e05f5a2d"
},
{
"id":"ev_e73541851aeb2dcc2caf",
"object":"obj_c78144b0b422f79531d614fd",
"actor":"the owner",
"at":"2026-10-09T06:49:14.272Z",
"ledger":"worker:misc-sheets:933f1d1a-b50a-4504-81a9-98a8218b7579:2",
"version":"34524ca537d0"
}
],
"dry_run": [
{
"id":"ev_0bd0c946b282bc1f835c",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"the owner",
"at":"2026-10-09T06:55:57.776Z",
"ledger":"worker:misc-sheets:bb4fb6e4-caa0-4c86-99ad-a7bd0595a612:2",
"version":"dbbc98c59c39"
},
{
"id":"ev_d3ba7d723c538fb2d9e2",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"the owner",
"at":"2026-10-09T06:49:15.884Z",
"ledger":"worker:misc-sheets:e0aba1f9-60bf-475b-a169-a25c998a4a41:2",
"version":"5b844c9bb3b9"
}
],
"executed": [
{
"id":"ev_3ae1c812905a25ef981a",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"the owner",
"at":"2026-10-09T06:55:56.603Z",
"ledger":"worker:misc-sheets:aac7e55f-d6ac-4ef9-8831-3257c2ee49e1:15",
"version":"ddb97c529860"
},
{
"id":"ev_2f9bacb67d27ba900cc8",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"the owner",
"at":"2026-10-09T06:55:51.990Z",
"ledger":"worker:misc-sheets:87700cd4-5728-45b3-8207-3e42702e6bb6:15",
"version":"2f50eb2e6723"
}
],
"observed": [
{
"id":"ev_e60c45e579f9a9dba975",
"object":"obj_0d4432a7d566c0872f27e2ec",
"actor":"the owner",
"at":"2026-10-09T06:55:56.573Z",
"ledger":"worker:misc-sheets:aac7e55f-d6ac-4ef9-8831-3257c2ee49e1:14",
"version":"bb37a689596d"
},
{
"id":"ev_931bce95dfa9d346272a",
"object":"obj_483635a9343e0b3becfc4671",
"actor":"the owner",
"at":"2026-10-09T06:55:51.959Z",
"ledger":"worker:misc-sheets:87700cd4-5728-45b3-8207-3e42702e6bb6:14",
"version":"39558ba9564e"
}
],
"activated": [
{
"id":"ev_0a154ccc436f3c1fc642",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"the owner",
"at":"2026-10-09T06:55:47.644Z",
"ledger":"worker:misc-sheets:af0a7c67-e9cd-4205-adf1-d5225849f6ea:2",
"version":"eba1f0f1d0e1"
}
],
"revised": [
{
"id":"ev_45ae00363d8fe7365855",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"Space tester (member) (person:p_n5cauqlp)",
"at":"2026-10-09T06:55:12.011Z",
"ledger":"worker:misc-sheets:6d5c1ad0-927e-4bfd-abdc-aa2f98e878d2:2",
"version":"d706bfb4560e"
}
],
"commented": [
{
"id":"ev_c765f9164af94749f9b3",
"object":"obj_6df3cb08754fb4008f21b3be",
"actor":"the owner",
"at":"2026-10-09T06:53:49.075Z",
"ledger":"worker:misc-sheets:9c36b4f8-a678-4fc4-b83c-c460e8acaf5a:1",
"version":"7d04c5d00c69"
}
],
"interpreted": [
{
"id":"ev_5841b0f0120c118c53fa",
"object":"obj_6df3cb08754fb4008f21b3be",
"actor":"the owner",
"at":"2026-10-09T06:53:47.095Z",
"ledger":"worker:misc-sheets:328db083-7e14-45c2-b43e-c65bb4e64c62:2",
"version":"ace451bc5aaa"
},
{
"id":"ev_b7ce4cd8fa4ebbb76f09",
"object":"obj_6df3cb08754fb4008f21b3be",
"actor":"the owner",
"at":"2026-10-09T06:46:53.830Z",
"ledger":"worker:misc-sheets:0cafed52-f128-452c-814e-6213ff33927f:11",
"version":"cb9e074a8c89"
}
],
"proposed": [
{
"id":"ev_df78d50820177291216e",
"object":"obj_7056be75b84c01cb84dcb14e",
"actor":"the owner",
"at":"2026-10-09T06:48:05.047Z",
"ledger":"worker:misc-sheets:3dde5558-62a6-4980-8eaf-5ec61c9ca800:2",
"version":""
},
{
"id":"ev_14f28b926e2dc0334fb0",
"object":"obj_6df3cb08754fb4008f21b3be",
"actor":"the owner",
"at":"2026-10-09T06:48:00.907Z",
"ledger":"worker:misc-sheets:8a1a7567-e8a8-44e0-9a91-537b09cd1eda:2",
"version":"fd89bd0e49bd"
}
],
"expanded": [
{
"id":"ev_9ed138fd41bc02a1b757",
"object":"obj_9a90225482d0d81ed742c2c1",
"actor":"the owner",
"at":"2026-10-09T06:47:37.163Z",
"ledger":"worker:misc-sheets:b900a8da-3688-498e-8399-91411e34f34c:23",
"version":"63f0726541dc"
}
],
"resolved": [
{
"id":"ev_a45a520a4402a17d61f0",
"object":"obj_6df3cb08754fb4008f21b3be",
"actor":"the owner",
"at":"2026-10-09T06:46:53.392Z",
"ledger":"worker:misc-sheets:0cafed52-f128-452c-814e-6213ff33927f:8",
"version":"e45a9b47f90c"
}
],
"ingested": [
{
"id":"ev_ec4ab243e035414d5c8d",
"object":"obj_6df3cb08754fb4008f21b3be",
"actor":"the owner",
"at":"2026-10-09T06:46:53.024Z",
"ledger":"worker:misc-sheets:0cafed52-f128-452c-814e-6213ff33927f:5",
"version":""
}
]
},
"executions": {
"ingest":"ex_d807cba10e24dd415231f02ba956",
"resolve":"ex_b88f6460728fe2b33e22f4ce9db2",
"interpret":"ex_f0a63b467224c58e054129995ba6",
"expand":"ex_9b8e26b9739578ac38596ee52a04",
"variants":"ex_a2d76ea53cc3ebe947287e9f0c27",
"policy_create":"ex_c9c135efd9b67b203290a090f102"
},
"proof_of_two_people": {
"what_this_is":"two separately authenticated humans and the deterministic Space agent on one canonical object graph; every step below is a live call the build answered, with its receipt",
"human_a": {
"who":"the owner",
"how":"the build key (x-sheets-key), the same credential the Console uses",
"viewer": {
"kind":"owner",
"id":"owner"
}
},
"human_b": {
"who":"Space tester (member)",
"person_id":"p_n5cauqlp",
"how":"a person record added by the owner (POST /api/people), a one-time invitation link made with link_only (not sent by the build; the owner hands it over), signed in by POST to that link: an ms_session cookie of their own",
"role_in_space":"editor",
"ops_granted": [
"read",
"analysis.confirm",
"variant.review",
"policy.revise",
"policy.dryrun",
"space.layout"
],
"viewer": {
"kind":"person",
"id":"p_n5cauqlp"
}
},
"agent": {
"id":"space-agent",
"kind":"deterministic",
"what_it_did": [
"resolved the advertiser from the landing page’s domain (ad.resolve, ex_b88f6460728fe2b33e22f4ce9db2)",
"wrote the analysis (ad.interpret, ex_f0a63b467224c58e054129995ba6)",
"read the watched pages twice on the owner’s press and recorded two observations (policy.check, ex_ab497a2a77c3836604e285791bef and ex_553e759661f9df0a7058c3663d5f)"
],
"model_calls":0
},
"canonical_object": {
"id":"obj_a0e0c09160cba70ca3d08e8a",
"type":"analysis",
"link":"https://miscsubjects.com/live/o/obj_a0e0c09160cba70ca3d08e8a"
},
"sequence": [
{
"step":1,
"who":"human B",
"read":"GET /live/s/obj_c78144b0b422f79531d614fd.json with their own session",
"saw": {
"me": {
"kind":"person",
"id":"p_n5cauqlp",
"role":"editor"
},
"objects":20,
"ops_offered_on_the_analysis": [
"analysis.confirm"
],
"ops_offered_on_the_rule": [
"policy.revise",
"policy.dryrun"
]
}
},
{
"step":2,
"who":"human A",
"read":"the analysis before any change",
"version":"282e5745512c",
"hook":"extracted",
"offer":"extracted"
},
{
"step":3,
"who":"human B",
"did":"POST /live/s/…/reply {object_id: the analysis, text: \"confirm hook\"}: read as analysis.confirm",
"execution":"ex_37f26ece904e17986804b8cd92ed",
"event":"corrected (ledger worker:misc-sheets:3898e13f-f8dd-4b4f-8825-c21f5b5cfacf:2)"
},
{
"step":4,
"who":"human B",
"did":"POST /live/objects/obj_a0e0c09160cba70ca3d08e8a/ops/analysis.confirm {field: offer, decision: reject, note: the figures $7, $48, $1.2 are per-task costs in the page text, not the offer the ad makes}",
"execution":"ex_38822e0a85530807918a15c452b3",
"version_after":"405141e27e3c",
"event":"corrected (ledger worker:misc-sheets:4b398714-831e-4a6e-8438-0ea9aaf5233e:2)",
"kept":"the earlier value in the statement’s history (history: 1)"
},
{
"step":5,
"who":"human A",
"read":"the same object, GET /api/blocks/objects/obj_a0e0c09160cba70ca3d08e8a",
"version":"405141e27e3c",
"hook":"confirmed by Space tester (member) (person)",
"offer":"rejected by Space tester (member) (person)",
"note":"the version the owner reads is the version B wrote; nothing was re-created"
},
{
"step":6,
"who":"human B",
"did":"POST /live/objects/obj_7056be75b84c01cb84dcb14e/ops/policy.activate",
"answer": {
"http":403,
"state":"refused",
"code":"not_granted",
"reason":"You were not granted policy.activate on this object.",
"ledger":"worker:misc-sheets:479bf1e6-a483-40cc-8113-53f34f0c8683:1"
},
"event":"denied"
},
{
"step":7,
"who":"human B",
"did":"POST …/ops/policy.revise {text: \"Only generate alternatives if they sell a product we also sell.\"}",
"execution":"ex_1d468782c0c102ca521558b791e6",
"result": {
"version":2,
"diff": {
"added": [
{
"condition": {
"field":"product.overlap",
"op":"eq",
"value":true
}
}
],
"changed": [
{
"gate":"variants_only_if_conditions",
"from":null,
"to":true
}
],
"removed": []
}
},
"event":"revised (ledger worker:misc-sheets:6d5c1ad0-927e-4bfd-abdc-aa2f98e878d2:2)"
},
{
"step":8,
"who":"human B",
"did":"reply to the ad: \"Find their other public pages and offers\" (read as competitor.expand)",
"answer":"refused: You were not granted competitor.expand on this object.",
"event":"denied (ledger worker:misc-sheets:5267144c-1c0f-40ee-a7e4-353e6291a656:1)"
},
{
"step":9,
"who":"human A",
"did":"policy.activate (ex_dcebeaccfb4860c8112159fbfc18): active, next check 2026-10-09T09:00:00Z; then policy.check twice on the owner’s press",
"observations": [
"obj_0d4432a7d566c0872f27e2ec: nothing changed on 4 watched pages (first reading)",
"obj_483635a9343e0b3becfc4671: nothing changed on 4 watched pages"
],
"dry_run_after": {
"facts": {
"ad.count":1,
"page.changed":false,
"offer.changed":false,
"product.overlap":false
},
"holds":false,
"would_run": []
}
},
{
"step":10,
"who":"a visitor (no sign-in)",
"before_sharing": {
"GET /live/s/….json":403,
"POST analysis.confirm":403
},
"after_the_owner_shared": {
"GET /live/s/….json": {
"me": {
"kind":"guest",
"role":"viewer"
},
"objects":22
},
"POST analysis.confirm (field cta, confirm)":"succeeded, ex_9c2394eaa1d2218376b771751def: the owner granted * [read, analysis.confirm] on this one object",
"POST policy.dryrun": {
"http":403,
"code":"not_granted",
"ledger":"worker:misc-sheets:be93ec73-ddde-438e-ae45-f6ec8c96fa71:1"
}
}
}
],
"where_the_decision_is_made":"src/blocks.js authorize(), on the server, from the object’s grants and the viewer the request proves (key, session cookie, or a visitor cookie); the widgets only draw what the server offers",
"synchronization":"polling: every surface reads the graph again every 20 seconds (not while someone types) and at once after any widget on the page changes an object; no push",
"not_claimed":"cryptographic non-repudiation. Each event carries the actor the server resolved and a Ledger row id; the Ledger is append-only on the build, not independently witnessed."
},
"real_and_not": {
"real_fetches":"every page in this Space was read by the build’s WEB_FETCH tool and kept with its sha256; the asset is the landing page as fetched",
"simulated_data":"none",
"the_ad":"the landing page’s own presentation of itself (og:title, og:description, og:image, its first call to action), labelled so on the ad object; a Meta Ad Library address is accepted but its content cannot be read from the cloud (Meta answers the cloud browser with a bot challenge; recorded on the asset when it happens)",
"model_calls_in_the_pipeline":0,
"inferences":"none made: the two statements that need one (positioning, audience) are left open and say so; the model path is the owner’s chat, where this object travels with the words",
"variant_images":"not rendered: variant.render is wired to Workers AI and runs only on the owner’s press; the variants here are objects with a headline, a prompt, a hypothesis and a review state",
"public_search_results":"stand in for an ad library and are labelled as such on the research object"
},
"where_it_stops": [
"Meta Ad Library pages are not readable from the cloud (bot challenge); the Mac scraper opens the owner’s own Chrome and was not run",
"two of nine statements of the analysis are open (positioning, audience): they need a model, which is not run here",
"the rule’s clock is read as UTC; the scheduled pass runs only when the switch spaces-monitor is on (off by default)",
"edge access scope is recorded (members) but every edge of a readable space is shown; per-edge disclosure is not enforced yet",
"the example base metrics were typed by the owner as an example and are labelled so on each metric; the site’s view counts are the only measured numbers here",
"Meta Ads tools of this build do not reach a service today (META_ADS_ACCOUNTS answers “no way that reaches a service”), so no ad-account metric is imported",
"offers: the competitor’s pricing page carries no prices in its HTML, so 0 products were made; the figures the landing page carries were extracted as “offer” and a person rejected the statement"
],
"metrics": [
{
"id":"obj_703c8f227a4d26ce8189b16d",
"kind":"impressions",
"value":24000,
"unit":"count",
"source":"the owner",
"as_of":"2026-10-09"
},
{
"id":"obj_521c887780e256ba0e20603d",
"kind":"aov",
"value":120,
"unit":"usd",
"source":"the owner",
"as_of":"2026-10-09"
},
{
"id":"obj_332b9793a08894e0960fbee4",
"kind":"cvr",
"value":2.1,
"unit":"%",
"source":"the owner",
"as_of":"2026-10-09"
},
{
"id":"obj_64db4ae4077be0adf1f80933",
"kind":"cpc",
"value":0.62,
"unit":"usd",
"source":"the owner",
"as_of":"2026-10-09"
},
{
"id":"obj_91e223a6e3f1de4e26a3fd2c",
"kind":"ctr",
"value":1.8,
"unit":"%",
"source":"the owner",
"as_of":"2026-10-09"
},
{
"id":"obj_eb8f2e4919af3e8cd354fc5d",
"kind":"views",
"value":19,
"unit":"count",
"source":"this build’s own view counter (POST /api/view) for live-business-spaces, software-objects-proof, directory-ledger-pattern, unified-dispatch-macaroon-atoms, atomic-tooling, task-payload-provenance-article",
"as_of":"2026-10-09"
}
],
"projections": [
{
"id":"obj_d7c69283387ec013341a60c1",
"name":"If CTR +10%, CPC -5%",
"version":2,
"assumptions": {
"aov": {
"op":"pct",
"value":15
},
"cpc": {
"op":"pct",
"value":-5
},
"ctr": {
"op":"pct",
"value":10
}
},
"missing": [],
"rows": [
{
"kind":"impressions",
"base":24000,
"after":24000,
"delta_pct":0,
"from":"measured"
},
{
"kind":"ctr",
"base":1.8,
"after":1.9800000000000002,
"delta_pct":10.000000000000009,
"from":"assumed"
},
{
"kind":"views",
"base":0,
"after":0,
"delta_pct":null,
"from":"measured"
},
{
"kind":"clicks",
"base":432,
"after":475.20000000000005,
"delta_pct":10.00000000000001,
"from":"derived: impressions × CTR"
},
{
"kind":"visits",
"base":432,
"after":475.20000000000005,
"delta_pct":10.00000000000001,
"from":"derived: clicks (when no visits are measured)"
},
{
"kind":"cpc",
"base":0.62,
"after":0.589,
"delta_pct":-5.000000000000004,
"from":"assumed"
},
{
"kind":"spend",
"base":267.84,
"after":279.8928,
"delta_pct":4.500000000000018,
"from":"derived: clicks × CPC"
},
{
"kind":"cpm",
"base":11.16,
"after":11.6622,
"delta_pct":4.500000000000002,
"from":"derived: spend ÷ impressions × 1000"
},
{
"kind":"cpa",
"base":29.52380952380952,
"after":28.047619047619044,
"delta_pct":-4.999999999999996,
"from":"derived: spend ÷ conversions"
},
{
"kind":"aov",
"base":120,
"after":138,
"delta_pct":15,
"from":"assumed"
},
{
"kind":"cvr",
"base":2.1,
"after":2.1,
"delta_pct":0,
"from":"measured"
},
{
"kind":"conversions",
"base":9.072000000000001,
"after":9.979200000000002,
"delta_pct":10.000000000000012,
"from":"derived: clicks × CVR"
},
{
"kind":"revenue",
"base":1088.64,
"after":1377.1296000000002,
"delta_pct":26.500000000000007,
"from":"derived: conversions × AOV"
},
{
"kind":"roas",
"base":4.064516129032259,
"after":4.920203735144312,
"delta_pct":21.052631578947345,
"from":"derived: revenue ÷ spend"
}
]
}
],
"needs": [
{
"id":"obj_980ef64bae4d5966100ac782",
"key":"ad_spend",
"what":"Their ad spend and targeting",
"how":"research",
"status":"open"
}
],
"issues": [
{
"id":"obj_6b730c2a0cdf739a96d8029d",
"task":"T-0312",
"title":"Confirm or reject the two open statements of the Relevance AI analysis",
"state":"open",
"about": [
"obj_a0e0c09160cba70ca3d08e8a"
]
}
],
"views": [
"overview",
"board",
"graph",
"table",
"timeline",
"rules",
"metrics",
"execution",
"chat",
"the object’s own page /live/o/<id>",
"the chat reply",
"this article"
],
"views_from_sentences": {
"table of pages by competitor sorted by date": {
"id":"obj_4911352ddcc7b01553572476",
"kind":"table",
"understood": [
"a table of pages",
"sorted by date, desc",
"grouped by competitor"
],
"unparsed": [],
"matches":24,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_4911352ddcc7b01553572476",
"execution":"ex_67cdeec34f3ebafc058d3bf73a00",
"error":null
},
"chart article views": {
"id":"obj_dc16723c982dbc11fd68eae4",
"kind":"chart",
"understood": [
"a chart of the measured metrics",
"the Page views series"
],
"unparsed": [],
"matches":null,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_dc16723c982dbc11fd68eae4",
"execution":"ex_9e15d3142e42dd56f9c7111b22cd",
"error":null
},
"compare us and Relevance AI": {
"id":"obj_625b9169c712162a0f178c85",
"kind":"compare",
"understood": [
"side by side: us and relevance ai"
],
"unparsed": [],
"matches":null,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_625b9169c712162a0f178c85",
"execution":"ex_aabcf2975999877efb10840319bf",
"error":null
},
"levers CTR CPC AOV": {
"id":"obj_a6b1f4e9b1f1ce9b1ae124ea",
"kind":"levers",
"understood": [
"levers for CTR, CPC, Average order value"
],
"unparsed": [],
"matches":null,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_a6b1f4e9b1f1ce9b1ae124ea",
"execution":"ex_cd9bf9a70748bc2f0b466413e1fd",
"error":null
},
"apps": {
"id":"obj_e3bd4204f00ec949992bdd85",
"kind":"apps",
"understood": [
"the Apps view: every object drawn as the app it lives in"
],
"unparsed": [],
"matches":123,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_e3bd4204f00ec949992bdd85",
"execution":"ex_8b3f1957eb73b4c7dae0b252557b",
"error":null
},
"what is missing to compare AOV": {
"id":"obj_7690d8d161349b5ed503aeb6",
"kind":"needs",
"understood": [
"what is not known to compare Average order value"
],
"unparsed": [],
"matches":null,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_7690d8d161349b5ed503aeb6",
"execution":"ex_b00df7a48781b3076492220f58d2",
"error":null
},
"if CTR up 10% and CPC down 5%, show the funnel": {
"id":"obj_8b27114b26c56eb4473b3b6a",
"kind":"funnel",
"understood": [
"the funnel after CTR +10%, CPC -5%"
],
"unparsed": [],
"matches":null,
"projection":"obj_8590b54b7d5345b3d03217e8",
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_8b27114b26c56eb4473b3b6a",
"execution":"ex_5180a89f889c7b15fb0089db5630",
"error":null
},
"show creatives by CTR": {
"id":"obj_1c1dad5f09682456d2e66cbb",
"kind":"table",
"understood": [
"a table of variants",
"sorted by CTR, high to low (an object that carries no CTR shows UNKNOWN)"
],
"unparsed": [],
"matches":3,
"projection":null,
"link":"/live/s/obj_c78144b0b422f79531d614fd?view=view:obj_1c1dad5f09682456d2e66cbb",
"execution":"ex_890c922cddd236e3c54708ed71ac",
"error":null
},
"make 3 variants please": {
"id":null,
"kind":null,
"understood":null,
"unparsed":null,
"matches":null,
"projection":null,
"link":null,
"execution":"ex_57d8bd7059eaccb74b58d16c1cb1",
"error":"not read as a view. A view starts with show, table, board, chart, graph, list, compare, levers, arrange, group, filter, sort, apps, “what is missing to compare …” or “if …, show the funnel”"
}
},
"view_grammar":"show | table | board | chart | graph | list <what> [by <field>] [where <field> <op> <value>] [sorted by <field> [asc|desc]] [as funnel|bars|lines|table|board] · compare <a> and <b> · levers <kinds> · arrange | group | filter | sort … · apps · what is missing to compare <metric> · if <scenario>, show the funnel"
},
"card_fields": [
"counts.objects",
"counts.edges",
"counts.events",
"the_competitor.name",
"the_rule.version"
],
"ontology": {
"above": [
"software-objects-proof"
],
"below": [],
"adjacent": [
"directory-ledger-pattern",
"unified-dispatch-macaroon-atoms",
"atomic-tooling"
],
"nested": [],
"show_position":false
},
"body": [
"Drop in a competitor ad. The Space turns it into objects, links them, and keeps watching.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"board",
"height":640
},
"## The same graph as a dashboard\n\nThe overview is the graph again, as numbered panels: the input, the analysis, the competitor, its pages, the variants, the graph, a direct reply field, the differences beside us (unknowns are said, never filled), what is not known (each a need a person or an agent can take), the rules as switches, the metrics, the timeline, the apps the engine registers, the people and the agent. Every panel reads the same ids.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"overview",
"height":900
},
"## One address in, a graph out\n\nThe address `https://relevanceai.com/` was dropped into the Space on 2026-10-09. The build fetched the page itself (sha256 `ade5239277240a02…`, 301339 bytes) and made these objects, each with its own id and its own page:\n\n- **asset** `obj_6d0c383a011ce1d35583e74b`: the original, with its hash and the exact fetch\n- **ad** `obj_6df3cb08754fb4008f21b3be`: what the page presents as its ad (headline, text, image, call to action); labelled as such\n- **competitor** `obj_9a90225482d0d81ed742c2c1`: Relevance AI, resolved from the landing page’s domain, never from a name alone; the evidence is on the edge\n- **pages** ×6 and **research** ×1: the competitor’s public pages (each with its hash) and public search results, labelled as search results\n- **analysis** `obj_a0e0c09160cba70ca3d08e8a`: nine statements, each extracted with its source, confirmed, rejected or left open\n- **claims** ×2: sentences the page makes, each an object the ad `contains_claim`\n- **hypotheses** ×3 and **variants** ×3: each variant `derived_from` the ad and `tests` one hypothesis\n- **rule** `obj_7056be75b84c01cb84dcb14e`: a sentence turned into a structure the runtime enforces\n\nEvery line on the board is a typed edge with its evidence, author, time and version. Click one.",
"## The ad, as an object",
{
"type":"object",
"id":"obj_6df3cb08754fb4008f21b3be",
"app":"ad"
},
"## Correct one statement\n\nThe analysis is not a paragraph. It is nine statements, each marked **extracted** (with the words it was read from), **confirmed**, **rejected** or **open**. Anyone reading this page may confirm, correct or reject a statement here; the change is recorded with who made it, the earlier value is kept, and the same statement changes on the board above and on the object’s own page.",
{
"type":"object",
"id":"obj_a0e0c09160cba70ca3d08e8a",
"app":"analysis"
},
"## The same object from a reply\n\nA reply addresses the object by its id, not a recap of a chat. The Space’s agent reads the words: *find their other public pages*, *make three variants*, *check this competitor daily* are operations; anything else is kept as a note on the object. What the reply made appears in the Space, not only here.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"chat",
"focus":"obj_6df3cb08754fb4008f21b3be",
"height":620
},
"## The rule\n\nThe sentence “Check this competitor daily. When a new offer appears, compare it with ours and propose three alternatives. Do not publish anything without approval. Only generate alternatives if they sell a product we also sell.” became version 1 of a rule: every day at 09:00; if offer.changed eq true and product.overlap eq true; then observe, compare_offer, propose_variants ×3; gates: approval before publish; approval before send; budget: max_fetches 4, model_calls 0, max_variants 3, max_runs_per_day 1. A dry run reads what the Space knows and spends nothing. A new sentence makes a new version with its difference against the one before. Activation and checks are gated by role; this build never publishes or sends from a rule.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"rules",
"height":560
},
"## Numbers, and a projection over them\n\nA metric is a number with a source and a date; this page invents none. Below, two kinds are real to this build: the views of six articles of this site (its own counter), and a base the owner typed in as an example, labelled as such. A projection is an object: the sentence “if CTR goes up 10% and CPC down 5%” became assumptions; the funnel was recomputed with the formulas written in the object (clicks = impressions × CTR, spend = clicks × CPC, conversions = clicks × CVR, revenue = conversions × AOV); every row names its measured source, its formula or its assumption; rows that need a metric nobody measured stay empty and say so. Toggle Traffic, Spend, Conversion or Market; transpose rows and columns; open the JSON under the view; reply to the projection to change an assumption (a new version, with its history).",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"metrics",
"height":620
},
"## Views from a sentence\n\nA view is a saved definition over the same objects: what to show, grouped by which field, filtered how, sorted how, drawn as a table, a board, a chart, a funnel or the apps. It is made by a sentence through one grammar (`show | table | board | chart | graph | list <what> [by <field>] [where <field> <op> <value>] [sorted by <field>] [as funnel|bars|lines]` · `compare <a> and <b>` · `levers <kinds>` · `arrange | group | filter | sort …` · `apps` · `what is missing to compare <metric>` · `if <scenario>, show the funnel`). What the grammar does not understand is said back, never guessed: “show the flibbertigibbet by moon phase” answers “not understood: by moon phase / flibbertigibbet”. A model saying any of this says it through the same grammar; no model was called to make these.\n\nThe sentences below were typed into this Space on 2026-10-09; each saved view, its id and what the grammar read are in `facts.views_from_sentences`. The levers on the Metrics view above are sliders: each writes a new projection version when it settles, and two moved together write one version; a projected row opens into the intermediate variables and the formula that produced it, each input marked measured, assumed, derived or missing.",
"**“table of pages by competitor sorted by date”**",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"view:obj_4911352ddcc7b01553572476",
"height":520
},
"**“chart article views”** — the six article view counts this build measures itself, each bar naming its source.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"view:obj_dc16723c982dbc11fd68eae4",
"height":480
},
"**“if CTR up 10% and CPC down 5%, show the funnel”** — the projection is made from the scenario, then the funnel rows open into their inputs.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"view:obj_8b27114b26c56eb4473b3b6a",
"height":620
},
"**“compare us and Relevance AI”**",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"view:obj_625b9169c712162a0f178c85",
"height":520
},
"**“apps”** — every object of the Space that lives in an app, drawn as that app: the issue as a Reminders card, the pages as Safari, the screenshots as Photos.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"view:obj_e3bd4204f00ec949992bdd85",
"height":640
},
"**The canvas** — two or more views side by side, each with its own width, saved with the Space.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"canvas",
"height":640
},
"## Issues on the same objects\n\nAn issue is a task of this build (its Tasks tab), opened from the Space with a zero-context spec composed from the object it is about, attached to the Space and linked to that object with an `about` edge. The board of issues is the Tasks list by state.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"execution",
"height":520
},
"## What happened, in order\n\nEvery ingest, resolution, interpretation, correction, proposal, dry run, activation and refusal is an event with its actor, the object’s version and its Ledger row.",
{
"type":"space",
"id":"obj_c78144b0b422f79531d614fd",
"view":"timeline",
"height":520
},
"## Two people, one object, the server deciding",
{
"type":"json",
"ref":"#facts.proof_of_two_people",
"title":"the proof: two signed-in people, one object, a change seen by both, a refusal",
"open_depth":2
},
"## What is real here, and what is not",
{
"type":"json",
"ref":"#facts.real_and_not",
"title":"real fetches, simulated data, model calls",
"open_depth":1
},
{
"type":"json",
"ref":"#facts.where_it_stops",
"title":"where it stops",
"open_depth":1
},
"## The object graph behind the page\n\nA Space is itself an object (type `space`, adapter `spaces`). Its members are grants on the space and on everything it contains; a role (viewer, editor, approver) is a list of operations. Objects live in `ms.app_objects`; edges in `ms.app_edges` (source, target, type, evidence, author, version, access); events in `ms.app_events`, each with a Ledger row. Every operation runs through the live-objects engine: authorized on the server for this viewer, written as an intent first, receipted. The graph is read at `GET /live/s/<space id>.json` by whoever opens it, under their own permissions.",
{
"type":"json",
"ref":"#facts.counts",
"title":"the graph today",
"open_depth":2
},
{
"type":"json",
"ref":"#facts.edges_by_type",
"title":"edges by type",
"open_depth":1
},
{
"type":"json",
"ref":"#facts.events",
"title":"events, by kind, with Ledger rows",
"open_depth":1
},
"![[software-objects-proof|the objects this is built on]]"
],
"sources": [
{
"id":"s1",
"url":"https://relevanceai.com/",
"title":"Specialist AI Agents for Every Task | Relevance AI",
"publisher":"Relevance AI",
"source_type":"company_site",
"quote":"Specialist AI Agents for Every Task",
"quote_sha256":"8aa2a25315fde91102db6ba032244c0ee617e1113e3de643e64af897d1f4c15c",
"http_status":200,
"snapshot_sha256":"ade5239277240a029f857f5718cdecdc658c2b42fbf6acb7a4c3bdf6581a5080",
"snapshot_bytes":301339,
"quote_found":true,
"verified_at":"2026-10-09T11:28:33.084Z"
},
{
"id":"s2",
"url":"https://relevanceai.com/enterprise",
"title":"Enterprise AI Agents | Relevance AI",
"publisher":"Relevance AI",
"source_type":"company_site",
"quote":"Enterprise AI Agents",
"quote_sha256":"ea88d748bb0627c35d09209b3a3010b5ff3494fb1788f5fb931564d20a424e18",
"http_status":200,
"snapshot_sha256":"e034b5d9945ada2440c4ac85ca3811b80900fee5988b686682f9c8bbd161add9",
"snapshot_bytes":242932,
"quote_found":true,
"verified_at":"2026-10-09T11:28:33.089Z"
},
{
"id":"s3",
"url":"https://miscsubjects.com/live/s/obj_c78144b0b422f79531d614fd.json",
"title":"the Space, as JSON, read without a key",
"publisher":"miscsubjects.com",
"source_type":"primary_record",
"quote":"Competitive intelligence",
"quote_sha256":"37a354b0ad68bb410fbd7858147e347c4f4b69958fdc33fe25479129a1f6c046",
"http_status":200,
"snapshot_sha256":"f655c39ee1d96936c34c179ffdee5c46a8cea62e20ae06ae60eca75741f7a712",
"snapshot_bytes":546470,
"quote_found":true,
"verified_at":"2026-10-09T11:28:33.436Z"
}
],
"comments": [],
"type":"article",
"created":"2026-10-09T06:57:08.023Z",
"updated":"2026-10-09T11:28:33.436Z",
"version":10,
"hash":"0ff1ab814fc730b170ffcb36711c101c62fa93b6379e9c16fc7ec7701c978e81",
"styling": {
"scope":"shared",
"how":"no value on the page is set by hand: each comes from this theme, and each theme value becomes the CSS variable named beside it. PATCH https://miscsubjects.com/api/jarticles/theme changes them.",
"widget_style": {
"what":"every widget may carry style: {…}; these are the only keys it reads",
"rotate":"degrees, a number",
"offset_x":"pixels, a number",
"offset_y":"pixels, a number",
"pulse":"true to pulse"
},
"theme": {
"color": {
"page_bg":"#f5f5f7",
"sheet_bg":"#ffffff",
"ink":"#1d1d1f",
"soft":"#424245",
"muted":"#6e6e73",
"dim":"#86868b",
"line":"#d2d2d7",
"accent":"#0071e3",
"accent_soft":"rgba(0,113,227,0.08)",
"link":"#0066cc",
"selection":"rgba(0,113,227,0.18)",
"widget_bg":"#f5f5f7",
"widget_border":"#d2d2d7",
"card_bg":"#ffffff",
"code_bg":"#f5f5f7",
"json_bg":"#161b19",
"json_ink":"#e6e1d4",
"json_key":"#7fd3c3",
"json_string":"#e9c46a",
"json_number":"#f4a261",
"json_bool":"#c792ea",
"json_null":"#928d7e",
"json_punct":"#6b6759",
"json_bar":"#0e1512",
"kind_badge_bg":"#0e6e60",
"kind_badge_ink":"#ffffff"
},
"color_dark": {
"page_bg":"#000000",
"sheet_bg":"#1d1d1f",
"ink":"#f5f5f7",
"soft":"#d2d2d7",
"muted":"#a1a1a6",
"dim":"#86868b",
"line":"#424245",
"accent":"#2997ff",
"accent_soft":"rgba(41,151,255,0.12)",
"link":"#2997ff",
"selection":"rgba(41,151,255,0.25)",
"widget_bg":"#2c2c2e",
"widget_border":"#424245",
"card_bg":"#1d1d1f",
"code_bg":"#2c2c2e",
"json_bg":"#0a0d0c",
"json_ink":"#e6e1d4",
"json_key":"#7fd3c3",
"json_string":"#e9c46a",
"json_number":"#f4a261",
"json_bool":"#c792ea",
"json_null":"#757062",
"json_punct":"#5d665f",
"json_bar":"#060807",
"kind_badge_bg":"#4fc1ab",
"kind_badge_ink":"#0b0f0d"
},
"font": {
"display":"-apple-system,BlinkMacSystemFont,\"SF Pro Display\",\"SF Pro Text\",\"Helvetica Neue\",Helvetica,Arial,sans-serif",
"body":"-apple-system,BlinkMacSystemFont,\"SF Pro Display\",\"SF Pro Text\",\"Helvetica Neue\",Helvetica,Arial,sans-serif",
"mono":"ui-monospace,SFMono-Regular,\"SF Mono\",Menlo,Consolas,monospace",
"google_fonts":""
},
"size": {
"body":"17px",
"small":"17px",
"eye":"17px",
"lead":"17px",
"h1":"52px",
"h2":"34px",
"h3":"22px",
"byline":"17px",
"h1_measure":"none",
"code":"17px",
"json":"17px",
"widget_title":"17px"
},
"leading": {
"body":"1.7",
"head":"1.18",
"lead":"1.5",
"json":"1.6"
},
"weight": {
"display":"700",
"head":"700",
"body":"400",
"strong":"700",
"eye":"700"
},
"tracking": {
"eye":"0.14em",
"display":"-0.03em",
"body":"-0.006em",
"badge":"0.08em"
},
"layout": {
"sheet_width":"1000px",
"text_width":"880px",
"gutter":"16px",
"sheet_padding":"56px 64px 64px",
"sheet_padding_mobile":"24px 18px 40px",
"sheet_margin":"28px auto 72px",
"block_gap":"1.15em",
"widget_gap":"28px",
"grid_gap":"14px"
},
"border": {
"sheet":"0px solid",
"sheet_radius":"30px",
"widget":"1px solid",
"widget_radius":"24px",
"card_radius":"24px",
"badge_radius":"999px",
"rule":"1px solid",
"code_radius":"18px",
"accent_bar":"3px"
},
"shadow": {
"sheet":"0 1px 1px rgba(0,0,0,.02),0 6px 26px rgba(0,0,0,.05)",
"widget":"none",
"card_hover":"0 10px 28px -16px rgba(28,27,23,0.35)"
},
"motion": {
"hover":"160ms ease"
}
}
},
"page":"https://miscsubjects.com/a/live-business-spaces",
"json":"https://miscsubjects.com/a/live-business-spaces.json",
"markdown":"https://miscsubjects.com/a/live-business-spaces.md",
"theme":"https://miscsubjects.com/article-theme.json",
"calls": [
{
"what":"this article as a page (no key)",
"method":"GET",
"url":"https://miscsubjects.com/a/live-business-spaces",
"headers": {}
},
{
"what":"this article as its object (no key)",
"method":"GET",
"url":"https://miscsubjects.com/a/live-business-spaces.json",
"headers": {}
},
{
"what":"this article as Markdown (no key)",
"method":"GET",
"url":"https://miscsubjects.com/a/live-business-spaces.md",
"headers": {}
},
{
"what":"the object with its hash, drafts too",
"method":"GET",
"url":"https://miscsubjects.com/api/jarticles/live-business-spaces",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)"
}
},
{
"what":"one piece and that piece’s hash",
"method":"GET",
"url":"https://miscsubjects.com/api/jarticles/live-business-spaces/at?path=facts",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)"
}
},
{
"what":"grep: every key or value that matches, with the path to change it",
"method":"GET",
"url":"https://miscsubjects.com/api/jarticles/live-business-spaces/grep?q=owner",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)"
}
},
{
"what":"change pieces (set, merge, delete, append, insert, move); if-match is the hash",
"method":"PATCH",
"url":"https://miscsubjects.com/api/jarticles/live-business-spaces",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)",
"content-type":"application/json",
"if-match":"<hash from the read>"
},
"body": {
"ops": [
{
"op":"set",
"path":"facts.example",
"value":1
}
]
}
},
{
"what":"replace the whole object",
"method":"PUT",
"url":"https://miscsubjects.com/api/jarticles/live-business-spaces",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)",
"content-type":"application/json",
"if-match":"<hash>"
},
"body": {
"slug":"live-business-spaces",
"title":"…"
}
},
{
"what":"fetch every source, hash what came back and check the quote is in it",
"method":"POST",
"url":"https://miscsubjects.com/api/jarticles/live-business-spaces/verify",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)",
"content-type":"application/json"
},
"body": {}
},
{
"what":"an Obsidian-style command",
"method":"POST",
"url":"https://miscsubjects.com/api/jarticles/cli",
"headers": {
"x-sheets-key":"$(cat /Users/cyrusmassoumi/.misc-sheets-key)",
"content-type":"application/json"
},
"body": {
"command":"backlinks",
"file":"live-business-spaces"
}
}
]
}