Hub · Connectors Integrations / Claude · Desktop · Code · Connectors

Ask Claude what you decided. Get what you wrote.

Claude is unmatched at reading what's in front of it. The 200K-token context window, the thoughtful tool use, the careful citations. But on Monday you write a decision; on Wednesday Claude is blank — the model sees enormous context, none of which is your context. The decision log lives in a Notion page Claude can't read, a Slack thread that scrolled away, a half-finished doc nobody opens.

Knovya is the part Claude reads. Anthropic-certified install — your archive walks with you across every Claude window, every session, every model version. Same memory, same toolkit, no re-explaining who you are or what you're working on.

34 tools · 5 families
3 install paths · ~90 sec
Listed in Anthropic's Connectors Directory
The install moment

Three Claudes. Three doorways. Same archive.

Claude Desktop, Claude Code, and the Anthropic Connectors Directory — each gets the same Knovya. Pick the path that matches your setup; the archive is the same on the other side.

Path I

Claude Desktop

Edit claude_desktop_config.json. Knovya appears in the connector menu next launch. The snippet above is the whole config.

Path II

Claude Code

From your terminal: claude mcp add knovya https://mcp.knovya.com/mcp. OAuth opens, Claude Code registers Knovya, /mcp lists 34 tools.

Path III

Connectors Directory

One-click from Anthropic's Connectors Directory. No JSON paste — pick Knovya, OAuth, done.

Behind the door

Same brain, 34 tools, organized for the conversation.

Claude reaches into Discovery first when you ask "what did we decide about X last quarter?" Read tools fire when it cites a customer interview. Write tools save the chat as a structured decision log — with your confirmation.

  • 0 Discovery 7 tools

    Where am I, what's here. Read-only orientation tools Claude uses first to check whether you've already written about the topic — before guessing from training data.

    pingworkspacehomepersonacontextsearchschema
  • 1 Read 7 tools

    Open the note. Trace a link. Pull an attachment. When Claude cites an ADR, a customer interview, or last quarter's retro — these are the tools doing the reaching.

    readexperiencememoryfoldershistorylinksattachments
  • 2 Write 4 tools

    Capture, edit, organize, archive. Save the conversation as a structured decision entry, tag it, file it. Destructive ops surface a confirmation step before changing your archive.

    writeeditorganizedelete
  • 3 Mesh 11 tools

    When more than one Claude is in the room. Claude Desktop hands off to Claude Code, or pings ChatGPT in another tab — presence, channels, shared attention, consensus voting.

    presencechannelscoordinateattentionthoughtscommitteeagentseventsmesh_adminpipelinenotifications
  • 5 Transform 5 tools

    Shape the archive. Convert a chat into a structured ADR. Apply your team's template. Export to Markdown for review. Where Knovya itself becomes a tool Claude can wield.

    aitemplatesexportimportshare
The problem

Claude reads 200,000 tokens of context. None of them, by default, are your context.

On Monday you reason through onboarding with Claude — three approaches, the trade-offs, the customer signal that pushed you to one. The conversation is sharp. You close the tab.

On Wednesday you open Claude again and ask what you decided about onboarding. The model gives you a careful, plausible answer — generated from training data, not from your Monday. The decision log lives in a Notion page Claude can't read, a Slack thread that scrolled away, a half-finished doc nobody opens. The intelligence is unmatched. The memory is somewhere else.

Knovya is the part Claude reads. The Monday session lives on; Wednesday opens to it already there. Same archive, every conversation, every Claude.

↳ §5 Lineage Anthropic's path to portable knowledge, in five steps

Lineage

How Claude got portable knowledge.

Anthropic walked the protocol forward. Five steps from chat-only AI to a connector layer that travels.

  1. 2022

    Anthropic ships the first Claude.

    A chat AI built around explicit values — helpful, harmless, honest. Brilliant in any single conversation, blank in the next. The intelligence is here; the memory isn't.

  2. 2023

    The long-context era arrives.

    Context windows leap — from a few thousand tokens to a hundred and beyond. Claude can finally read a whole document at once. But the documents still live elsewhere, and the model can't reach them without copy-paste.

  3. 2024

    Anthropic ships the Model Context Protocol.

    An open spec for AI ↔ tools. Any client can talk to any server, the same way browsers talk to web servers. Anthropic publishes the protocol, the SDKs, and the first reference servers. Knowledge becomes portable.

  4. 2025

    The Anthropic Connectors Directory launches.

    A curated catalog of MCP servers — vetted, OAuth-ready, one-click. Claude Desktop and Claude Code both speak the protocol. The connector layer is no longer a developer ritual; it's a product surface.

  5. 2026 · today

    Knovya joins the Connectors Directory.

    Filesystem reads files. GitHub reads commits. Knovya reads your reasoning — decisions, retros, customer interviews, ADRs. Claude finally walks into the conversation knowing what you wrote.

First mover

Other MCPs reach for files and tickets. Knovya reaches for your reasoning.

Filesystem

Anthropic's own reference MCP. Reads and writes local files outside the workspace.

Local files

GitHub

PRs, issues, commits, releases — Claude reads code history without leaving the chat.

Repo ops

Slack

Search messages and channels. Pull a quote from yesterday's discussion thread.

Chat history

Sentry

Error events, stack traces, release health — Claude debugs with the actual error in hand.

Error data

Linear

Issues, sprint state, project rollups — the work your team has already broken into tickets.

Tasks

Knovya

Your knowledge graph — decisions, retros, customer interviews, the architecture you wrote — as first-class context. The reasoning that doesn't fit in a file or a ticket.

Decision memory

They stack. Most Claude users connect three or four MCPs at once — Filesystem for local reads, GitHub for code history, Slack for context, Knovya for the reasoning. Different layers of the same archive. Claude picks the right tool per question.

Surfaces

Four moments. One archive.

Where the connector actually shows up — the morning conversation, the terminal, the config, the directory.

Claude Desktop · 7:14 AM i.

"What did we decide about onboarding?"

You · 7:14 AM
"What did we decide about onboarding last quarter?"
↳ Claude reaches into your archive
knovya_search(query="onboarding decision")
knovya_experience(topic="onboarding flow Q4")
↳ 4 notes · 2 successful precedents · 1 cautionary
✓ Cites the ADR, the customer interviews, and the retro you wrote.

Same question, different answer. Claude returns what you actually wrote — not what it guesses from training data.

Claude Code · terminal ii.

One CLI, then 34 tools.

$ claude mcp add knovya https://mcp.knovya.com/mcp
→ Opening browser for OAuth…
✓ Authorized · 34 tools registered.
$ claude
› /mcp
filesystem · github · knovya (34)
› refactor /api/auth using past decisions
knovya_search(query="auth strategy decisions")
✓ Patch drafted citing ADR-014.

Claude Code's /mcp command lists every connected server. Knovya shows up alongside Filesystem and GitHub.

claude_desktop_config.json iii.

The Desktop config, one paste.

{
  "mcpServers": {
    "knovya": {
      "url": "https://mcp.knovya.com/mcp",
      "transport": "http"
    },
    "filesystem": {
      "command": "npx",
      "args": ["@modelcontextprotocol/server-filesystem"]
    }
  }
}

OAuth handles sign-in on first launch — no API key in the file, no token in your shell. Restart Claude Desktop and the connector menu has Knovya in it.

Anthropic Connectors Directory iv.

Listed. Vetted. One-click.

Featured connectors vetted
Filesystem official
GitHub official
Knovya connect
Sentry connect

Skip the JSON. From the directory, hit Connect to Knovya, OAuth in, done. Same archive — fewer steps.

Bonded with

Claude isn't alone. Same archive, four other doorways.

Knovya is the part that doesn't change between tools. Whichever AI catches your next thought, it's reading from the same archive Claude reads.

ChatGPT

The midnight draft. Past sessions, saved quotes, half-finished threads — already there.

Walk through ChatGPT
Cursor

When the same archive needs to live inside the IDE. Composer cites your ADRs.

Walk through Cursor
Gemini

Long-context research. Synthesis written back to your archive instead of trapped in a chat.

Walk through Gemini
Knovya API

REST + webhooks. For the integration we haven't pre-built yet — Slack bots, Zapier, custom agents.

Walk through the API
Free 50 MCP calls / month — no credit card
Pro · $15/mo 5,000 calls · encrypted notes
Team · $25/seat Unlimited · shared archive
See pricing
Questions, in advance

Claude & Knovya — before you ask.

What is Claude MCP?

Claude MCP is Claude's support for the Model Context Protocol — an open spec from Anthropic that lets Claude read from and write to external tools. Knovya is one of those tools.

Once added, Claude can search your archive, read your decision log, and cite past notes inline. It works the same in Claude Desktop, Claude Code, and the Anthropic Connectors Directory.

How do I add Knovya to Claude Desktop?

Open claude_desktop_config.json — on macOS at ~/Library/Application Support/Claude/, on Windows at %APPDATA%\Claude\. Add a knovya entry under mcpServers with the MCP URL.

Restart Claude Desktop. Knovya appears in the connector menu, OAuth handles sign-in. About 90 seconds end-to-end.

How do I use Knovya with Claude Code?

In your terminal, run claude mcp add knovya https://mcp.knovya.com/mcp. The CLI registers the server, opens a browser for OAuth, and adds it to your Claude Code settings.

From the next session, /mcp lists Knovya's 34 tools and Claude Code can call any of them mid-task — search the archive while debugging, save the patch reasoning when you ship.

What is the best MCP server for Claude?

It depends on the work. For local files, Filesystem. For repo and PR ops, GitHub. For chat history, Slack. Knovya is the only MCP that brings your knowledge graph — decisions, retros, customer interviews, the reasoning your team wrote down — into Claude as first-class context.

Most Claude power users connect three or four MCPs at once; Knovya is the layer that holds the reasoning between the files.

Is Knovya in Anthropic's Connectors Directory?

Yes. Knovya is listed in Anthropic's Connectors Directory — the curated catalog of MCP servers Anthropic vouches for. From the directory you can connect Knovya to Claude in one click; OAuth 2.1 with PKCE handles the sign-in.

No JSON paste required if you'd rather skip that.

How is Knovya different from Claude Projects?

Claude Projects scope a conversation to a set of files — useful for keeping context tight inside one chat. Knovya is a structured archive that lives outside any single conversation, queryable with hybrid search and ranked by NoteRank.

Projects organize files for the model. Knovya holds the reasoning behind them — and it's reachable from Claude, ChatGPT, Cursor, Gemini, Copilot, and Windsurf, not just one product.

Is there a free MCP server for Claude?

Yes. Knovya Free includes 50 MCP calls per month — enough to try the full memory layer with no credit card, on Claude Desktop or Claude Code. Pro at $15/month lifts the cap to 5,000 calls and adds end-to-end encryption. Team makes it unlimited at $25 per seat.

Open Claude in the morning.
Knovya in the room.

Desktop in the morning, Code at the desk, Connectors Directory the day you onboard a teammate. Whichever Claude you walk in with, the archive you wrote is already there.

Get connected Browse all 7 connectors

Or jump to the install paths, then see pricing.

Listed in Anthropic's Connectors Directory. Free Knovya tier · 14-day Pro trial · No credit card to start.