ContextStream MCP Server
The context platform for AI agents: code search, team decisions, and grounded project context.
At a glance
Not tested yet- GitHub stars
- 47
- Weekly downloads
- 3,286
- Last commit
- 2026-11-19
- Licence
- MIT
- Transport
- stdio
- Auth
- api-key
npx -y @contextstream/mcp-server
What it does
ContextStream MCP Server connects your code, docs, and conversations so AI agents can find relevant files, recall saved decisions, and build on past work across sessions and tools. It is aimed at developers who want their agent to know the decisions behind their code, not just the code itself. It works with MCP clients such as Claude Code, Cursor, Codex, GitHub Copilot, Gemini CLI, Zed, and others. A free tier offers 10,000 monthly credits with no credit card required.

Tools it gives your agent
studio: ask for a film, product explainer, image, sound or music in plain words, and ContextStudio's producer plans the pieces from what your team knows.plan: returns the pieces and one price and spends nothing.make: spends at most the plan's price, and only after you agree.status: reports each piece while it waits.result: shows images inline and links video and sound.revise: starts from a made image or sound with avariation,revisionorresize.stop,library,brand,products: additional studio tools.
Quick start
Run the install script for your platform, then follow onboarding to connect your project and supported editor. Restart your editor after setup.
curl -fsSL https://contextstream.io/scripts/mcp.sh | bash
irm https://contextstream.io/scripts/mcp.ps1 | iex
Add this to .mcp.json in your project root (Claude Code) or .cursor/mcp.json (Cursor):
{
"mcpServers": {
"contextstream-mcp-server": {
"command": "npx",
"args": [
"-y",
"@contextstream/mcp-server"
],
"env": {
"CONTEXTSTREAM_API_KEY": "YOUR_CONTEXTSTREAM_API_KEY",
"CONTEXTSTREAM_API_URL": "YOUR_CONTEXTSTREAM_API_URL"
}
}
}
}Replace each YOUR_ value with your own, and delete the variables you do not use.
Environment variables
CONTEXTSTREAM_API_KEY: the key thatsetup --yesreads, or that you can save withcontextstream-mcp configure --api-key-stdin.CONTEXTSTREAM_API_URL: not explained in the sources.
Example prompts
- Use ContextStream to find the files relevant to my next change. Cite the sources and retrieve any saved project decisions that should guide the work.
- Save this decision and its reason so I can retrieve it in a new session.
- Plan a product explainer film from my connected product and show me the price before making anything.
What to watch out for
- Indexing sends eligible source contents to ContextStream for hosted search.
- Transcript saving and local Git metadata capture are on by default and can be turned off.
- Review data handling and controls before connecting sensitive projects.
- A hosted connection alone does not sync your local checkout.
makespends credits, and a product film takes 13 to 21 minutes.- Signed result links are valid about 15 minutes.
Not tested by tools4agents yet. Read the source repository before you install it.
Similar tools
Everything
The official test server. Exercises every part of the protocol, so it is the fastest way to check your setup works at all.
Filesystem
Gives an agent scoped read and write access to directories you name explicitly. The usual first MCP server people install.
Git
Read, search, and manipulate local Git repositories. Lets an agent inspect history and diffs without shelling out.
Sequential Thinking
Gives a model a scratchpad for multi step reasoning, letting it revise earlier steps instead of committing to a first answer.