The knowledge + memory substrate for AI agents.

TPS Report sits at the intersection of two territories: markdown-to-static publish tools and AI brain / Graph RAG platforms. Most products do one. We give both audiences one codified structure — humans browse it, agents execute on it — with RBAC on every page that readers and agent retrieval both obey. Monetize with it if you want.

Obsidian + MCP native The Retrieval Contract RBAC for humans & agents

Loading platform features...

Comparison dimensions

What we measure TPS Report against — and what most competitors in our territory cannot combine.

  • Reasoning-grade retrieval Hybrid graph + vector: multi-hop traversal, centrality scoring, and prerequisite chains — not naive chunk search.
  • The Retrieval Contract Every doc declares what it's for — and what it's not for. A versioned metadata contract enforced at authoring and at sync, so agents never misfire.
  • Reachable everywhere — plugins & MCP Browser-native, an Obsidian sync plugin for vault-native authors, and a native MCP server that exposes your corpus as tools to any MCP agent — Claude, Cursor, whatever ships next.
  • Bidirectional editing (cloud + local) Edit online or in Obsidian; sync keeps local vault and cloud corpus aligned both ways.
  • Access control — RBAC for humans & agents Role-based access on every report, section, and page — and agent retrieval obeys the same rules as readers. Monetization is optional on top: the same RBAC powers a membership portal when you want one.

How TPS Report compares

Five dimensions that matter for a knowledge + memory substrate: reasoning-grade retrieval, the Retrieval Contract, reach (plugins & MCP), bidirectional editing, and RBAC for humans & agents. Most territory products cover one or two — not all five.

DimensionTPS ReportTypical publish / doc / RAG toolsAdvantage
Reasoning-grade retrievalHybrid graph + vector retrieval — multi-hop traversal, PageRank-style centrality, and prerequisite chains ("read this first") on your codified corpus.Flat embedding search (RAG APIs) or no retrieval at all (publish tools) — nearest-chunk matching with no relationship awareness.Agents follow real relationships between documents, not embedding proximity.
The Retrieval ContractEvery document carries a versioned metadata contract — hyde_questions, retrieval_hint, defers_to, canonical_for, and a mandatory "do not use for…" clause — enforced at authoring and at sync.Basic YAML tags or none. Frontmatter is cosmetic, not an agent routing or canonical-source contract.Agents know what a document is for — and what it isn't — so they don't over-retrieve or misfire.
Reachable everywhere — plugins & MCPAuthor online or in Obsidian via a bidirectional sync plugin. A native MCP server exposes your corpus as tools to any MCP-compatible agent — not captive to one runtime.Obsidian-native only with no agent layer, cloud-only with no vault path, or RAG APIs with no plugin/MCP surface at all.Author where you already work; let any agent — ours or not — reach the same substrate.
Bidirectional editing (cloud + local)Edit in the online editor or Obsidian; sync flows both ways so cloud and local stay the same source of truth before publish.One-way vault → static site, or cloud-only with no local editor. Edits in one place do not round-trip.No fork between "where you write" and "what you publish and index."
Access control — RBAC for humans & agentsRole-based access on every report, section, and page — and agent retrieval enforces the same rules as readers. Monetize on top when you want: the same RBAC powers public teasers, member unlocks, and a signup funnel.Public static sites, basic password gates, or RAG indexes that leak everything to every query — access control is bolted on, and agents bypass it.One access model governs readers and agents alike — govern first, monetize if desired.

Reasoning-grade retrieval

TPS Report
Hybrid graph + vector retrieval — multi-hop traversal, PageRank-style centrality, and prerequisite chains ("read this first") on your codified corpus.
Typical publish / doc / RAG tools
Flat embedding search (RAG APIs) or no retrieval at all (publish tools) — nearest-chunk matching with no relationship awareness.
Advantage: Agents follow real relationships between documents, not embedding proximity.

The Retrieval Contract

TPS Report
Every document carries a versioned metadata contract — hyde_questions, retrieval_hint, defers_to, canonical_for, and a mandatory "do not use for…" clause — enforced at authoring and at sync.
Typical publish / doc / RAG tools
Basic YAML tags or none. Frontmatter is cosmetic, not an agent routing or canonical-source contract.
Advantage: Agents know what a document is for — and what it isn't — so they don't over-retrieve or misfire.

Reachable everywhere — plugins & MCP

TPS Report
Author online or in Obsidian via a bidirectional sync plugin. A native MCP server exposes your corpus as tools to any MCP-compatible agent — not captive to one runtime.
Typical publish / doc / RAG tools
Obsidian-native only with no agent layer, cloud-only with no vault path, or RAG APIs with no plugin/MCP surface at all.
Advantage: Author where you already work; let any agent — ours or not — reach the same substrate.

Bidirectional editing (cloud + local)

TPS Report
Edit in the online editor or Obsidian; sync flows both ways so cloud and local stay the same source of truth before publish.
Typical publish / doc / RAG tools
One-way vault → static site, or cloud-only with no local editor. Edits in one place do not round-trip.
Advantage: No fork between "where you write" and "what you publish and index."

Access control — RBAC for humans & agents

TPS Report
Role-based access on every report, section, and page — and agent retrieval enforces the same rules as readers. Monetize on top when you want: the same RBAC powers public teasers, member unlocks, and a signup funnel.
Typical publish / doc / RAG tools
Public static sites, basic password gates, or RAG indexes that leak everything to every query — access control is bolted on, and agents bypass it.
Advantage: One access model governs readers and agents alike — govern first, monetize if desired.

Products in our territory

Competitors mapped against the five dimensions above — most cover one or two, not all five.

Obsidian & markdown → static site

Digital gardens, vault-native publish, self-hosted static generators

Obsidian Publish

Hosted SaaS ($8/mo+) — select vault folders, graph view, wikilinks

Strength: Lowest friction for Obsidian users; native graph + publish UX

Gap vs TPS: No gated KB product layer, no agent-oriented metadata contract, no agent execution brain

Quartz

Open-source SSG — Obsidian-flavored MD, graph, backlinks, Git deploy

Strength: Free, battle-tested digital garden; upstream-safe customization

Gap vs TPS: Static site only — you build/host; no publish workflow, access control, or RAG layer

Foliate

Vault-native SSG — config in `.foliate/`, public vs published tiers

Strength: Flexible Publish alternative; stays inside vault; incremental builds

Gap vs TPS: Static output only; no online authoring, monetization, or agent retrieval

Kiln

Go binary SSG — Obsidian parity (canvas, callouts, wikilinks)

Strength: Single binary, fast builds, deploy anywhere

Gap vs TPS: Site generator, not a KB publishing + AI brain platform

Flowershow

Obsidian plugin + hosted publish — CSS/JSON theming

Strength: Cheaper than Publish; granular site control from vault

Gap vs TPS: Publish-focused; no structured agent metadata or execution substrate

Nuartz

Next.js + shadcn Obsidian publish stack (Quartz-inspired)

Strength: React ecosystem customization for digital gardens

Gap vs TPS: Developer SSG; not end-to-end KB + agent product

Astro / Hugo + MD

General SSG + markdown/Git — plugins for wikilinks

Strength: Maximum control, zero vendor lock-in, great performance

Gap vs TPS: Engineering-heavy; no built-in KB gating, metadata contract, or agent chat

Documentation & knowledge-base SaaS

Team docs, developer portals, wikis with optional AI Q&A

GitBook

Git + visual editor SaaS — spaces, AI search/Q&A, collaboration

Strength: Cross-functional docs; AI-assisted answers on published docs

Gap vs TPS: Docs portal, not codified graph metadata for agent routing; weak on Obsidian/vault path

Mintlify

Developer-docs SaaS — MDX, OpenAPI, AI assistant, MCP/llms.txt

Strength: Fast API docs, AI-native dev experience

Gap vs TPS: Dev-docs niche; not vertical KB products or agent-oriented metadata routing

Notion + AI

Workspace wiki — Q&A over pages, collaboration

Strength: Familiar editor; broad team adoption

Gap vs TPS: Flat page model; not publish-first KB with Graph RAG metadata contract

Confluence + Atlassian Intelligence

Enterprise wiki — Rovo/AI search across Atlassian stack

Strength: Enterprise permissions, Jira integration

Gap vs TPS: Internal wiki; not customer-facing gated KB + agent execution product

ReadMe / Redocly

API reference & developer hub platforms

Strength: Interactive API docs, try-it playgrounds

Gap vs TPS: API docs vertical; not general codified knowledge for agents

Docusaurus / Starlight

Open-source docs frameworks (Meta/Astro)

Strength: Self-hosted, customizable doc sites

Gap vs TPS: Static docs codebases; no managed publish + agent brain layer

Outline

Team wiki — self-host or cloud, markdown-friendly

Strength: Clean wiki UX, reasonable self-host story

Gap vs TPS: Collaboration wiki; not metadata-rich agent retrieval platform

Guru / Slite / Tettra

Team knowledge bases with AI search

Strength: Internal enablement, browser extensions, quick capture

Gap vs TPS: Internal KB tools; not publish + Graph RAG for external/agent products

AI brain, Graph RAG & agent memory

Context engines, enterprise search, temporal KG memory — ingest & retrieve

Glean

Enterprise Work AI — unified index, Enterprise Graph, agents, 100+ connectors

Strength: Permission-aware search at scale; org + personal graphs

Gap vs TPS: Indexes existing SaaS sprawl — you don’t publish structured KB products inside it

Dust

Multiplayer agent workspace — shared agents, company context

Strength: Team agent building and governance

Gap vs TPS: Agent orchestration layer; not author-publish-index your own codified corpus

Ragie

Context engine API — vector + keyword + summary indexes, MCP, entity extraction

Strength: Production RAG pipeline as a service; multi-tenant isolation

Gap vs TPS: Ingestion/retrieval infra — you bring unstructured docs; no publish workflow or metadata contract

Zep (Graphiti)

Temporal knowledge graph / Context Lake — bi-temporal facts, hybrid retrieval

Strength: Graph-first agent memory; invalidates stale facts over time

Gap vs TPS: Memory API — not a human-facing KB publish + gating product

Mem0

Memory layer for agents — extraction pipeline, optional graph (Mem0g)

Strength: Drop-in personalization; LangChain/LlamaIndex integrations

Gap vs TPS: Conversation memory — not structured report authoring + canonical routing

Vectara / Ragie-class RAG APIs

Managed RAG — chunk, embed, retrieve, rerank

Strength: Fast path to “chat with PDFs”

Gap vs TPS: File dump RAG; no defers_to, canonical_for, or publish layer

Onyx (formerly Danswer)

Open-source enterprise search + chat over connectors

Strength: Self-hostable Glean-like experience

Gap vs TPS: Connector search — not codified KB authoring + productized reports

Neo4j GraphRAG / LangGraph patterns

DIY graph + vector stacks

Strength: Full control for engineering teams

Gap vs TPS: Build-it-yourself; TPS Report productizes publish + metadata + retrieval

OpenAI / Anthropic file stores + assistants

Upload files → assistant with retrieval

Strength: Quick prototypes

Gap vs TPS: No graph metadata, gating, or KB-as-product workflow

Adjacent (partial overlap)

Worth knowing — touch publish OR retrieval, rarely both

Sanity / Contentful / headless CMS

Structured content APIs → frontends

Strength: Content modeling for apps and marketing sites

Gap vs TPS: CMS for apps — not agent-oriented Graph RAG metadata or chat on corpus

Intercom Fin / Zendesk AI

Support bots trained on help center articles

Strength: Customer support automation

Gap vs TPS: Support vertical; articles are help docs, not agent routing graphs

Custom GPTs / Copilot Studio

Upload knowledge → branded chatbot

Strength: Fast internal bots

Gap vs TPS: Closed garden; no publish, monetize, or metadata contract across reports

Ready to publish a KB product with an AI brain?

Codify your expertise, gate access, and let agents retrieve the right context — not just the nearest chunk.