Product / Manifest

The Manifest is the product.

Synorb turns source events into machine-readable intelligence objects. Every Manifest keeps the source, the structured record, the brief, and the extracted signals together with stable IDs and provenance.

Observed sources
10,000+
Streams
1,000+
Primitive types
3
Home domains
12

One source event. Three machine-native views.

A Manifest is the complete unit Synorb writes. Agents can inspect the canonical JSON, read a concise Brief, or act on individual Signals without losing lineage back to the original source.

3 Things CIOs Know That Every AI Founder Should Understand - Issue #24

AI's primary barrier to adoption is not technological models but how companies are structured and operate. Founders must focus on rebuilding workflows for AI, not just optimizing existing systems.

engineering-technology Mayfield Blog text high
4d ago
76 claims
1 stream
Highlights the engineering-technology domain with insights for AI founders from CIOs, distinct from the investment angle.
Manifest
id 1777525429698648000
{
  "manifest_id": "1777525429698648000",
  "record_id": "1777525429649909800",
  "stream_ids": ["17723038993580068"],
  "stream_names": ["mayfield"],
  "home_domain": "engineering-technology",
  "source": {
    "name": "Mayfield Blog",
    "media_format": "text",
    "claim_type": "publication",
    "published_date": "2026-04-30"
  },
  "entities": {
    "organizations": ["Mayfield", "OpenAI", "Microsoft"],
    "people": ["CIOs", "AI founders"],
    "topics": ["AI adoption", "workflow design", "enterprise AI"]
  },
  "outputs": {
    "signals_count": 76,
    "brief_id": "1777527738510498300",
    "record_version": 1
  },
  "domain_classification": {
    "home_domain": "engineering-technology",
    "cross_domains": [
      "economics-business-work",
      "society-law-government",
      "people-biography-history"
    ]
  },
  "provenance": {
    "source_url": "https://mayfield.com/...",
    "captured_at": "2026-04-30T05:42:48.748631+00:00",
    "matched_at": "2026-04-30T05:42:49.006112+00:00",
    "pipeline_version": "synorb-unified-story-v3",
    "schema_version": "manifest/1"
  },
  "lineage": {
    "source_channel_id": "mayfield-blog",
    "stream_id": "17723038993580068",
    "record_id": "1777525429649909800",
    "brief_id": "1777527738510498300",
    "signal_ids": [
      "1777527738714483700",
      "1777527738801269800",
      "1777527738893154300"
    ]
  },
  "delivery": {
    "formats": ["signal", "brief", "record"],
    "interfaces": ["REST", "MCP", "webhook", "S3"],
    "refresh_cadence": "continuous"
  }
}
Signal
{
  "signal_id": "1777527738714483700",
  "manifest_id": "1777525429698648000",
  "record_id": "1777525429649909800",
  "stream_id": "17723038993580068",
  "claim": "AI adoption is blocked by how companies are built, not by model capability.",
  "claim_type": "analysis",
  "evidence": "paraphrase",
  "confidence": "stated",
  "sentiment": "neutral",
  "significance": "high",
  "source": {
    "name": "Mayfield Blog",
    "source_type": "organization",
    "media_format": "text",
    "published_date": "2026-04-30"
  },
  "tags": {
    "organizations": ["Mayfield"],
    "people": ["CIOs", "AI founders"],
    "topics": ["AI adoption", "workflow design", "enterprise AI"]
  },
  "domain_classification": {
    "home_domain": "engineering-technology",
    "cross_domains": ["economics-business-work", "society-law-government", "people-biography-history"]
  },
  "evidence_ref": {
    "source_url": "https://mayfield.com/...",
    "quote_mode": "paraphrase",
    "source_span": "AI's primary barrier to adoption is not technological models..."
  }
}
Brief
{
  "brief_id": "1777527738510498300",
  "manifest_id": "1777525429698648000",
  "stream_name": "mayfield",
  "headline": "CIOs frame AI adoption as workflow reconstruction",
  "summary": "The article argues that enterprise AI value depends on rebuilding decision paths, operating models, and human workflows around machine capability.",
  "key_points": [
    "Enterprise AI adoption is framed as an organizational design problem rather than a model quality problem.",
    "Founders should build around workflow reconstruction, decision latency, and operating model change.",
    "The CIO perspective shifts the sales conversation from model performance to implementation path."
  ],
  "notable_quotes": [],
  "actionable_takeaways": [
    "Map the workflow before selling the AI layer.",
    "Show how the customer organization changes once the system is adopted.",
    "Treat change management as product surface area."
  ],
  "source_urls": ["https://mayfield.com/..."],
  "published_date": "2026-04-30",
  "home_domain": "engineering-technology"
}
Record
{
  "record_id": "1777525429649909800",
  "manifest_id": "1777525429698648000",
  "title": "3 Things CIOs Know That Every AI Founder Should Understand - Issue #24",
  "url": "https://mayfield.com/...",
  "source_name": "Mayfield Blog",
  "source_channel": "mayfield-blog",
  "source_type": "organization",
  "media_format": "text",
  "claim_type": "publication",
  "content": "AI's primary barrier to adoption is not technological models but how companies are structured and operate...",
  "published_date": "2026-04-30",
  "captured_at": "2026-04-30T05:42:48.748631+00:00",
  "entities": {
    "organizations": ["Mayfield", "OpenAI", "Microsoft"],
    "people": ["CIOs", "AI founders"],
    "places": [],
    "topics": ["AI adoption", "workflow design", "enterprise AI"]
  },
  "routing": {
    "stream_names": ["mayfield"],
    "home_domain": "engineering-technology",
    "cross_domains": ["economics-business-work", "society-law-government", "people-biography-history"]
  },
  "version": 1
}

Same event. Different jobs.

The format choice is not cosmetic. Signals are for reasoning and alerts. Briefs are for dashboards and RAG context. Records are for joins, lineage, replay, and warehouse-grade use.

Signal

Atomic assertions

Claims with evidence, confidence, sentiment, source, tags, and date. Built so agents can reason over the smallest useful unit.

Brief

Compact narrative

A source-aware summary that keeps the important claims together. Built for dashboards, digests, and retrieval-augmented systems.

Record

Canonical JSON

The structured content object with stable identifiers, provenance, versioning, source metadata, and machine-joinable fields.

What produces a Manifest?

Source Channels are the exact surfaces Synorb watches. Streams are the canonical rollups agents subscribe to. The Manifest is what gets written when those watched surfaces publish.

Source Channel

The publishing surface

Blogs, filings, podcasts, feeds, reports, social posts, data releases, research pages, and other surfaces.

google-cloud odd-lots-podcast bls-consumer-price-index
Stream

The canonical subscription

A saved query or entity rollup that bundles related channels under a durable object agents can follow.

Alphabet Odd Lots BLS CPI
Manifest

The machine object

Signal, Brief, and Record views with stable IDs, provenance, lineage, domain classification, and typed tags.

manifest_id record_id stream_ids

Streams start from three primitive shapes.

Synorb does not treat every source as a loose feed. Streams resolve to primitives that agents can reason about: organizations, people, and datasets.

Primitive / Organization

Organizations

Companies, labs, banks, governments, publishers, funds, universities, agencies, and public institutions.

Alphabet6,028 /168h
Stripe553 /168h
Federal Reserve620 /168h
Sequoia Capital258 /168h
Primitive / Person

People

Founders, operators, researchers, investors, executives, policymakers, creators, and domain specialists.

Andrej Karpathy812 /168h
Paul Graham210 /168h
Dwarkesh Patel505 /168h
Sarah Guo243 /168h
Primitive / Dataset

Datasets

Company filings, research reports, economic indicators, podcasts, corporate blogs, statistical releases, and more.

SEC EDGAR2,410 /168h
BLS CPI144 /168h
Weather Alerts390 /168h
FOMC Calendar116 /168h

A slice of the primitive inventory.

The matrix is intentionally broad: organizations, people, and datasets live together because agents need to join them together. A person can cite a company; a company can publish a data point; a dataset can move a market narrative.

Organization
Alphabet
6,028 /168h
Organization
Stripe
553 /168h
Organization
Nvidia
317 /168h
Organization
Goldman Sachs
229 /168h
Organization
KPMG
852 /168h
Organization
Databricks
366 /168h
Organization
Federal Reserve
620 /168h
Organization
Cleveland Fed
311 /168h
Organization
Sequoia Capital
258 /168h
Organization
Oaktree Capital
499 /168h
Organization
Salesforce
546 /168h
Organization
Cloudflare
397 /168h
Person
Andrej Karpathy
812 /168h
Person
Paul Graham
210 /168h
Person
Dwarkesh Patel
505 /168h
Person
Sarah Guo
243 /168h
Person
Wilfred Frost
585 /168h
Person
David Rubenstein
269 /168h
Person
Jensen Huang
178 /168h
Person
Jerome Powell
441 /168h
Person
Patrick O'Shaughnessy
329 /168h
Person
Marc Andreessen
198 /168h
Person
Dan Luu
117 /168h
Person
Julia Evans
94 /168h
Dataset
SEC EDGAR
2,410 /168h
Dataset
BLS CPI
144 /168h
Dataset
Employment Situation
304 /168h
Dataset
JOLTS
188 /168h
Dataset
Producer Price Index
221 /168h
Dataset
Weather Alerts
390 /168h
Dataset
Weather Forecast
264 /168h
Dataset
Sports Odds
506 /168h
Dataset
FOMC Calendar
116 /168h
Dataset
Earnings Calls
712 /168h
Dataset
Vehicle Data
86 /168h
Dataset
Auction Results
139 /168h

A Stream bundles every channel an entity publishes.

Alphabet is one Stream. It includes developer, research, cloud, AI, product, and corporate channels. An agent subscribes to the Stream and receives Manifests from every linked Source Channel.

Example Stream
Alphabet
home_domain: engineering-technology · source_type: organization · 14 source channels
Claims / 168h
6,028
Google Cloudgoogle-cloud
428
Google DeepMindgoogle-deepmind
900
AIgoogle-developers-ai
389
Cloudgoogle-developers-cloud
77
Mobilegoogle-developers-mobile
808
Webgoogle-developers-web
226
Channels
14
All Alphabet-published surfaces.
Claims / 168h
6,028
Atomic assertions extracted last 7 days.
Avg / Channel
430
Per-channel claim rate.
Resolution
<1 min
Capture-to-Manifest latency.

The top-level ontology for every Manifest.

Every Manifest has one home_domain: the primary place an agent should file it. Cross-domains preserve the secondary context, but the home domain keeps the graph stable.

economics-business-work
Capital, companies, labor, and strategy

Markets, management, financial institutions, employment, company updates, allocation, and business operations.

Examples: earnings, bank research, VC, podcasts, corporate news
engineering-technology
Software, systems, AI, and infrastructure

Engineering work, platforms, developer tools, AI research, cloud, semiconductors, cybersecurity, and technical strategy.

Examples: engineering blogs, AI labs, cloud channels
society-law-government
Policy, institutions, law, and civic systems

Public institutions, legislation, regulation, elections, courts, agencies, civic infrastructure, and government action.

Examples: Fed policy, regulators, government releases
health-medicine
Medicine, care, biology, and public health

Healthcare delivery, biotech, clinical research, public health, pharmaceuticals, medical devices, and life sciences.

Examples: hospital systems, FDA updates, biotech reports
life-environment
Climate, energy, ecology, and living systems

Environment, agriculture, energy transition, weather impacts, climate systems, conservation, and resource use.

Examples: weather alerts, energy reports, climate research
physical-sciences-mathematics
Formal and physical sciences

Physics, chemistry, mathematics, materials, measurement, scientific discovery, and quantitative research.

Examples: research papers, lab updates, science datasets
arts-culture-entertainment
Culture, media, sports, and creative work

Film, television, music, games, sports, cultural industries, media production, and entertainment markets.

Examples: sports odds, media companies, creator channels
people-biography-history
People and institutions over time

Careers, biographies, leadership changes, historical narratives, institutional memory, and individual influence.

Examples: person streams, executive moves, interviews
language-literature
Writing, language, and interpretation

Books, essays, rhetoric, translation, language systems, literary work, and communication as a domain.

Examples: essays, publishing, linguistic research
everyday-life-practical-knowledge
Practical life and applied know-how

Consumer behavior, work practices, education, how-to knowledge, local decisions, and usable everyday context.

Examples: guides, education, consumer updates
places-geography
Countries, cities, regions, and place

Geographic context, regional economies, urban systems, geopolitics, demographics, and physical place.

Examples: regional data, city reports, country analysis
universe-earth
Earth systems and planetary scale

Space, astronomy, planetary science, earth systems, geophysics, and phenomena beyond local human systems.

Examples: astronomy, geology, planetary datasets

See Full Schema · Free API Keys Included

Get the complete manifest schema reference and 1,000 free manifests/month.

Schema PDF + free credentials delivered instantly with 1,000 manifests/mo on monthly delivery.