DevBrain
DevBrain is a server that enables you to interact with developer-curated content through the following functionalities:
Search Content: Query for relevant developer articles, blogs, and code snippets based on specific search terms
Filter Results: Ground searches using optional comma-separated tags or keywords
Read Articles: Retrieve the full content of an article by providing its URL
Token Management: Get the current authentication token or set a new one
Allows running the DevBrain MCP server as a Docker container for containerized deployment and integration with Claude
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@DevBrain find articles about implementing authentication in Next.js"
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.
DevBrain MCP Server
Chat with your favorite newsletters (coding, tech, founder).
Audit
|
About
It is a newsletter-based MCP that searches for relevant code snippets, indie developer articles and blog posts so you don't have to hunt through generic web results again. Just ask LLM: "research on devbrain"
It's kind of like a web search, but specifically tuned for high-quality, developer-curated content. You can easily plug in your favorite newsletter to expand its knowledge base even further.
For example, when you are implementing feature "A", DevBrain can pull related articles that would serve as a solid reference and a foundation for your implementation.
Claude app | Goose app (tap on an image to open utube) |
DevBrain returns articles as short description + URL, you can then:
instruct LLM agent like
ClaudeorGooseto fetch full contents of the articles using provided URLsinstruct LLM to implement a feature based on all or selected articles
Try quickly with remote http-transport MCP
"devbrain": {
"type": "http",
"url": "https://devbrain.svenai.com/mcp"
},Related MCP server: Smart Coding MCP
Local Installation (remote http mcp recommended)
Via uv or uvx. Install uv and uvx (if not installed):
curl -LsSf https://astral.sh/uv/install.sh | shExample command to run MCP server in stdio mode:
uvx --python ">=3.10" --from devbrain devbrain-stdio-serverUse in Claude Code (or other coding agents)
https://docs.anthropic.com/en/docs/claude-code/mcp#installing-mcp-servers You can either add MCP to cc manually or reference tthe same .json file that Claude app uses.
Use in Claude
To add devbrain to Claude's config, edit the file:
~/Library/Application Support/Claude/claude_desktop_config.json
and insert devbrain to existing mcpServers block like so:
{
"mcpServers": {
"devbrain": {
"command": "uvx",
"args": [
"--python", ">=3.10",
"--force-reinstall",
"--from",
"devbrain",
"devbrain-stdio-server"
]
}
}
}Claude issues:
Somehow it fails to get the latest version even when OS has it installed. Forcing an update (at least once) is required for Claude app. This is done with
--force-reinstallarg.Claude is known to fail when working with
uvanduvxbinaries. See related: https://gist.github.com/gregelin/b90edaef851f86252c88ecc066c93719. If you encounter this error then run these commands in a Terminal:
sudo mkdir -p /usr/local/binsudo ln -s ~/.local/bin/uvx /usr/local/bin/uvxsudo ln -s ~/.local/bin/uv /usr/local/bin/uvand restart Claude.
Integration for Cline and other AI agents
Command to start DevBrain MCP in stdio mode:
uvx --python ">=3.10" --force-reinstall --from devbrain devbrain-stdio-serverand add this command to a config file of the AI agent (Cline or other).
Note that DevBrain requires Python 3.10+ support. Most systems have it installed. However VS Code (that Cline depends on) is shipped with Python 3.9. Use correct version of Python when running DevBrain MCP. A corrected version to launch DevBrain MCP looks like this:
uvx --python ">=3.10" --force-reinstall --from devbrain devbrain-stdio-serverDocker integration
You can run this MCP as a Docker container in STDIO mode. First build an image with build.sh. Then add a config to Claude like so:
{
"mcpServers": {
"devbrain": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"svenai/mcp-devbrain-stdio:latest"
]
}
}
}Test command to verify that docker container works correctly:
docker run -i --rm svenai/mcp-devbrain-stdio:latestLicense
This project is released under the MIT License and is developed by mimeCam as an open-source initiative.
Available Tools
2 toolsread_full_articleC
Returns the full content of an article identified by its URL.
Args: url: The URL of the article to read.
Returns: str: The full content of the article or an error message.
Inputs: url (required). Returns: result
retrieve_knowledgeA
Queries DevBrain (aka developers brain` system) and returns relevant information.
Args:
query: The question or ask to query for knowledge.
tags: Optional comma-separated list of tags (keywords) to filter or ground the search. (e.g.: ios, ios,SwiftUI, react-native, web, web,react, fullstack,react-native,flutter). Do not provide more than 3 words.
Returns: str: Helpful knowledge and context information from DevBrain (articles include title, short description and a URL to the full article to read it later).
Inputs: query (required), tags. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 2 tools
The two tools have completely distinct purposes: one fetches full article content from a specific URL, while the other queries a knowledge base with a search query and optional tags. There is no functional overlap or ambiguity between these operations.
Both tools follow a clear verb_noun pattern (read_full_article, retrieve_knowledge) with consistent snake_case formatting. The naming conventions are predictable and semantically appropriate for their functions.
With only 2 tools, the server feels severely under-scoped for a system described as a 'developer's brain' knowledge base. A comprehensive knowledge system would typically need more operations like searching, filtering, listing, or managing knowledge entries beyond just retrieving content and querying.
The tool surface is incomplete for a knowledge base system. While it provides retrieval of specific articles and knowledge queries, it lacks essential operations like listing available articles, searching by metadata, updating knowledge, or managing the knowledge base structure. This creates significant gaps for agent workflows.
Maintenance
Related MCP Connectors
Search independent software the big engines bury: indie apps, open-source repos, and dev tools.
Read-only search over the Founder Ventures library on bootstrapped SaaS, plus founder events.
One semantic search across your sites, Drive, Notion, email, files and Basecamp.
Search and get recommendations from your own saved catalog of developer tools and services.
Related MCP Servers
- AlicenseAqualityDmaintenanceProvides semantic search and access to Contraption Company blog posts and essays. Uses AI-powered embeddings to find relevant content and supports real-time updates via webhooks.31MIT
- AlicenseAqualityDmaintenanceProvides intelligent semantic code search using local AI embeddings, enabling natural language queries to find relevant code by meaning rather than exact keywords. Indexes codebases in the background with smart project detection and privacy-first local processing.626 npm200MIT
- AlicenseNot gradedqualityDmaintenanceEnables semantic search across indexed documents using vector embeddings. Index GitHub repositories and URLs to perform natural language queries with AI-enhanced contextual results.16 npmMIT
- AlicenseAqualityDmaintenanceSemantic search across Claude Code conversations. Hybrid vector + keyword search, fully local, background indexing.621 npm11MIT