How do I use Point MCP Server?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Point MCP Server Search for documentation on OAuth 2.0 PKCE" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

Point MCP Server

by mcdonaldsam

Overview Schema Source Related Servers Discussions Admin

Python

Remote

Point MCP Server

MCP server for the Point Knowledge API — verified, citable knowledge for AI coding assistants.

Point indexes curated technical documentation (RFCs, framework docs, standards, API references) and makes it searchable with hybrid search (BM25 + vector) and precise citations. This MCP server gives your AI assistant direct access to that knowledge.

Tools

Tool	Description	Tokens
`search`	Hybrid search with citations and relevance scores	~200/result
`get_document_toc`	Lightweight table of contents for a document	~50
`get_sections`	Load specific sections by chunk ID (max 50)	varies
`list_collections`	Browse or search available knowledge collections	~100/collection
`get_document_full`	Full markdown content of a document	varies (can be large)

Recommended workflow: search or list_collections to find content, then get_document_toc for structure, then get_sections for specific passages. Use get_document_full only when you need the complete text.

Related MCP server: mcp-rag-server

Prerequisites

Python 3.11+ installed
Point API key — get one free at pinchpoint.dev/point/keys

Installation

pip install point-mcp

Or install from source:

git clone https://github.com/mcdonaldsam/point-mcp.git
cd point-mcp
pip install -e .

Setup by IDE

Claude Code

Add to your Claude Code MCP settings (~/.claude/settings.json or project .claude/settings.json):

{
  "mcpServers": {
    "point": {
      "command": "point-mcp",
      "env": {
        "POINT_API_KEY": "your-api-key-here"
      }
    }
  }
}

Or add via CLI:

claude mcp add point -- point-mcp -e POINT_API_KEY=your-api-key-here

Cursor

Add to your Cursor MCP config (~/.cursor/mcp.json):

{
  "mcpServers": {
    "point": {
      "command": "point-mcp",
      "env": {
        "POINT_API_KEY": "your-api-key-here"
      }
    }
  }
}

Windsurf

Add to your Windsurf MCP config (~/.windsurf/mcp.json):

{
  "mcpServers": {
    "point": {
      "command": "point-mcp",
      "env": {
        "POINT_API_KEY": "your-api-key-here"
      }
    }
  }
}

VS Code (GitHub Copilot)

Add to your VS Code settings (.vscode/mcp.json in your project, or user settings):

{
  "servers": {
    "point": {
      "type": "stdio",
      "command": "point-mcp",
      "env": {
        "POINT_API_KEY": "your-api-key-here"
      }
    }
  }
}

Using uvx (no install needed)

If you have uv installed, you can run point-mcp without installing it globally:

{
  "mcpServers": {
    "point": {
      "command": "uvx",
      "args": ["point-mcp"],
      "env": {
        "POINT_API_KEY": "your-api-key-here"
      }
    }
  }
}

Manual / Other Tools

Any MCP client that supports stdio transport:

POINT_API_KEY=your-api-key-here point-mcp

Configuration

Environment Variable	Required	Default	Description
`POINT_API_KEY`	Yes	—	Your Point API key (get one)
`POINT_API_URL`	No	`https://point-api.pinchpoint.dev`	API base URL (for self-hosted or local dev)

Examples

Once configured, your AI assistant can use Point tools naturally:

"Search Point for how OAuth 2.0 PKCE works"

"What collections does Point have about cloud infrastructure?"

"Get the table of contents for document rfc-7636, then load sections 2 and 3"

The assistant will automatically use the appropriate tools and include citations in its responses.

Development

# Clone and install with dev dependencies
git clone https://github.com/mcdonaldsam/point-mcp.git
cd point-mcp
pip install -e ".[dev]"

# Run tests
pytest

# Run server locally
POINT_API_KEY=your-key point-mcp

License

MIT

Available Tools

5 tools

get_document_fullA

Get the full markdown content of a document.

WARNING: This can be large (thousands of tokens). Prefer get_document_toc + get_sections for targeted retrieval. Use this only when you need the complete text.

Name	Required	Description	Default
`doc_id`	Yes	Document ID (from search results or collection listings)

Output Schema

Name	Required	Description
`result`	Yes

get_document_tocA

Get a lightweight table of contents for a document (~50 tokens).

Use this to understand document structure before loading specific sections. Returns chunk IDs that you can pass to get_sections.

Name	Required	Description	Default
`doc_id`	Yes	Document ID (from search results or collection listings)

Output Schema

Name	Required	Description
`result`	Yes

get_sectionsA

Load specific document sections by their chunk IDs (max 50 per request).

Use after search or get_document_toc to retrieve full text of specific passages. Each section includes its text, heading path, word count, and citation.

Name	Required	Description	Default
`chunk_ids`	Yes	List of chunk IDs to load (get these from search results or TOC). Max 50 per request.

Output Schema

Name	Required	Description
`result`	Yes

list_collectionsA

Browse or search available knowledge collections.

Without a query, lists all collections with document counts. With a query, performs semantic search to find relevant collections.

Name	Required	Description	Default
`query`	No	Optional search query to find relevant collections (e.g. "machine learning frameworks"). Omit to list all.

Output Schema

Name	Required	Description
`result`	Yes

searchA

Search Point's verified knowledge base using hybrid search (BM25 + vector).

Returns ranked results with relevance scores, text excerpts, and citations. Each result includes a pre-formatted citation you can use directly.

Name	Required	Description
`query`	Yes	Natural language search query (e.g. "how does OAuth 2.0 PKCE work")
`collection`	No	Optional collection ID to search within (e.g. "rfc-ietf"). Omit to search all collections.
`doc_type`	No	Optional document type filter (e.g. "rfc", "standard", "guide")
`limit`	No	Max results to return (1-50, default 10). Use 3-5 for focused queries.

Output Schema

Name	Required	Description
`result`	Yes

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

5 tool updatesv0.1.0May 25, 2026
- First observedget_document_full
- First observedget_document_toc
- First observedget_sections
- First observedlist_collections
- First observedsearch

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A4.4/5.0

Scored 2026-05-25 19:30 across 5 tools

Disambiguation5/5

Each tool has a distinct purpose: full document retrieval, table of contents, section loading, collection listing, and search. No overlap or ambiguity.

Naming Consistency5/5

All tool names follow a consistent verb_noun pattern (get_document_full, get_document_toc, get_sections, list_collections, search). Even 'search' fits as a verb with an implied object.

Tool Count5/5

With 5 tools, the server is well-scoped for knowledge base retrieval. It is neither too sparse nor too heavy.

Completeness5/5

The tool set covers all necessary operations for a read-only knowledge base: browsing collections, searching, getting document structure, and retrieving full or partial content. No gaps for its stated purpose.

Maintenance

ActivityInactive

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

forkast-mcp-docsA
io.github.ForkastPredict
MCP server for querying Forkast documentation
Documentation Access Search Developer Tools
42
https://mcp.forkast.gg/mcp
mcpA
ai.llmse
Public MCP server for the LLM Search Engine
Search RAG Systems Developer Tools
102
https://llmse.ai/mcp
iknow
dev.iknow
Knowledge base MCP for AI agents on iknow.dev. Search, read, and maintain via OAuth.
Knowledge & Memory Search Documentation Access
1
https://iknow.dev/mcp
OnceAskA
io.github.rsb6061
MCP server for OnceAsk, the AI-native current-address layer for people and agents.
Customer Data Platforms Search Location Services
41
https://onceask.com/_api/mcp

Related MCP Servers

mcp-ragdocs
RAG Systems Vector Databases Documentation Access
hannesrudolph
A
license
Not graded
quality
D
maintenance
An MCP server implementation that provides tools for retrieving and processing documentation through vector search, enabling AI assistants to augment their responses with relevant documentation context
Updated a year ago (2025-07-18 06:59 UTC)
15 npm
263
MIT
mcp-rag-server
RAG Systems Search
kwanLeeFrmVi
A
license
Not graded
quality
D
maintenance
An MCP server that indexes documents and serves relevant context to LLMs via Retrieval Augmented Generation (RAG).
Updated 2 years ago (2025-04-18 17:10 UTC)
29 npm
36
MIT
api-docs-mcp
RAG Systems Documentation Access
akshaysadanand
A
license
Not graded
quality
D
maintenance
An MCP server that gives AI coding assistants access to up-to-date API documentation via RAG by crawling documentation sites, indexing them into a vector store, and enabling semantic queries.
Updated 3 months ago (2026-08-08 06:58 UTC)
MIT
accreta
Knowledge & Memory Search Documentation Access
francescofioredev
A
license
Not graded
quality
B
maintenance
MCP server that enables AI agents to search, fetch, and analyze a self-maintaining markdown knowledge base with provenance, drift detection, and canonical definitions.
Updated 12 days ago (2026-10-28 18:20 UTC)
1
MIT

Point MCP Server

Tools

Prerequisites

Installation

Setup by IDE

Claude Code

Cursor

Windsurf

VS Code (GitHub Copilot)

Using uvx (no install needed)

Manual / Other Tools

Configuration

Examples

Development

License

Available Tools

Output Schema

Output Schema

Output Schema

Output Schema

Output Schema

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

mcp-ragdocs

mcp-rag-server

api-docs-mcp

accreta