Skip to content
MCP server

Dibri

Answers drawn only from your approved documents, with sources, a score or a refusal

Install
Pricing
Freemium
Platforms
API, Web
Runs
Remotely
Version
1.0.0
Listing updated
1 day ago

About Dibri

Dibri is a freemium MCP server for knowledge and memory. It runs as a hosted remote server, so there is nothing to install locally.

Dibri is built around a refusal. An agent asks a question and gets an answer composed only from documents in one vault, with numbered source references, the exact passages used, and a 0 to 100 confidence score. If nothing verified covers the question it returns blocked: true so the agent can hand off to a person instead of guessing. The docs suggest treating a score below about 50 as a cue to escalate.

Vaults can be filled by uploads, website crawls with scheduled recrawls, or OneDrive and Google Drive, and an approval step can hold documents until the owner approves them. Each agent answers from exactly one vault. The answering model has no tools, retrieval is scoped to your organisation and vault, and documents that try to instruct an AI are held for review at ingestion.

Over MCP there is a single ask tool taking a question and optional conversation history, authenticated with an agent token. An analytics key swaps it for trust tools that expose usage, refusals and knowledge gaps.

Install

Run in a terminal

claude mcp add --transport http dibri https://dibri.ai/mcp

Add to Settings → Connectors → Add custom connector

https://dibri.ai/mcp
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "dibri": {
      "url": "https://dibri.ai/mcp"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "dibri": {
      "type": "http",
      "url": "https://dibri.ai/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "dibri": {
      "serverUrl": "https://dibri.ai/mcp"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "dibri": {
      "type": "http",
      "url": "https://dibri.ai/mcp"
    }
  }
}

Tools this server provides (1)

Taken from the project documentation; the running server may differ.

ask
Ask a question and get a sourced answer with confidence score or a refusal.

What it does

Verified answers
Answers built only from approved documents with source passages.
Confidence score
Each answer carries a 0 to 100 score, and 0 when blocked.
Refusals
Unsupported questions are blocked so the agent can escalate.
Vault sources
Upload files, crawl sites or connect OneDrive and Google Drive.
Injection defences
The answering model has no tools and retrieval is scoped in code.
Trust data
Read usage, refusals, knowledge gaps and staleness through an analytics key.

Good for

  • Let a support agent answer product questions from official documents
  • Route unanswerable questions to a human
  • Find which documents are missing from refusal analytics
  • Resell scoped agents to clients with their own vaults

Pricing

Freemium, from $29 a month

Free 100 answers/mo; Builder $29/mo; Growth $99/mo; Scale $299/mo

Prices change. Check dibri.ai before you buy.

Dibri FAQ

What does it cost?

Free gives 100 answers a month; Builder is $29, Growth $99 and Scale $299 a month.

Do refusals count toward the limit?

Yes. A refusal counts as an answer.

What happens at the limit?

The API returns HTTP 402 until the limit resets on the 1st or you upgrade.

Is my data used for training?

The site says its model provider's terms bar training and Dibri never trains on your content.

Is Dibri free?

Dibri has a free plan, with paid plans for more usage or features. Paid plans start at $29 a month.

How do I add Dibri to Claude Code?

Run this command in your terminal: claude mcp add --transport http dibri https://dibri.ai/mcp

What are the best alternatives to Dibri?

The closest alternatives to Dibri are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and note1.

Alternatives to Dibri

  1. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  2. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  3. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  4. Local-first long-term memory layer for AI agents, served over MCP

    Not checked yet
    • 5 tools
    • Remote
    • Open source
    • 1.3K GitHub stars
    • Knowledge and memory
  5. note1

    MCP server

    MCP server giving AI assistants search and summaries across your note1 meetings

    • 4 tools
    • Remote
    • Free trial
    • 1 GitHub star
    • Productivity