DevBrain
# DevBrain MCP Server
**Chat with your favorite newsletters** (coding, tech, founder).
# Audit
| <a href="https://glama.ai/mcp/servers/@mimeCam/mcp-devbrain-stdio">
<img width="380" height="200" src="https://glama.ai/mcp/servers/@mimeCam/mcp-devbrain-stdio/badge" alt="DevBrain MCP server" /></a> | [](https://mseep.ai/app/mimecam-mcp-devbrain-stdio) |
|:--------:|:--------:|
| | [](https://mseep.ai/app/121bc8fb-67e7-4d57-b953-2d30b91cdfb5) |
# 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 <topic> 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._
| <img width="400" alt="usage-claude" src="https://github.com/user-attachments/assets/f87b80ee-7829-43e8-9223-a02a38b4fd12" /> | [](https://youtu.be/7UFtKqI9CjQ) |
|:--------:|:--------:|
| 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 `Claude` or `Goose` to fetch full contents of the articles using provided URLs
- instruct LLM to implement a feature based on all or selected articles
## Try quickly with remote http-transport MCP
```json
"devbrain": {
"type": "http",
"url": "https://devbrain.svenai.com/mcp"
},
```
## Local Installation (remote http mcp recommended)
Via `uv` or `uvx`. Install `uv` and `uvx` (if not installed):
```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
```
Example command to run MCP server in `stdio` mode:
```bash
uvx --python ">=3.10" --from devbrain devbrain-stdio-server
```
## Use 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:
```json
{
"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-reinstall` arg.
- [Claude is known to fail](https://gist.github.com/gregelin/b90edaef851f86252c88ecc066c93719) when working with `uv` and `uvx` binaries. See related: https://gist.github.com/gregelin/b90edaef851f86252c88ecc066c93719. If you encounter this error then run these commands in a Terminal:
```bash
sudo mkdir -p /usr/local/bin
```
```bash
sudo ln -s ~/.local/bin/uvx /usr/local/bin/uvx
```
```bash
sudo ln -s ~/.local/bin/uv /usr/local/bin/uv
```
and restart Claude.
## Integration for Cline and other AI agents
Command to start DevBrain MCP in `stdio` mode:
```bash
uvx --python ">=3.10" --force-reinstall --from devbrain devbrain-stdio-server
```
and 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:
```bash
uvx --python ">=3.10" --force-reinstall --from devbrain devbrain-stdio-server
```
## Docker 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:
```json
{
"mcpServers": {
"devbrain": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"svenai/mcp-devbrain-stdio:latest"
]
}
}
}
```
Test command to verify that docker container works correctly:
```bash
docker run -i --rm svenai/mcp-devbrain-stdio:latest
```
## License
This project is released under the MIT License and is developed by mimeCam as an open-source initiative.
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.