neuron-explorer
Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results
Code Ocean MCP Server
Translation pending
Developer toolsPythonGitHub @codeoceanSelf-installOwner unconfirmed
Paste into Claude or ChatGPT for installation instructions suited to your environment.
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Installable projects can be verified without a deployment domain. Use GitHub sign-in for a personal repository, or the repository verification file above for an organization repository.
Ask about other verification options
Loading the full README. The saved excerpt is shown below.
Model Context Protocol (MCP) server for Code Ocean.
This MCP server provides tools to search, run, and release capsules and pipelines, and manage data assets.
• Prerequisites
• Code Ocean Platform Version Compatibility
• Installation
• Visual Studio Code
• Claude Desktop
• Cline
• Roo Code
• Cursor
• Windsurf
• Streamable HTTP Transport
• Local Testing
1. Install uv from Astral or the GitHub README
2. Install Python 3.10 or newer using uv python install 3.10 (or a more recent version)
3. Generate a Code Ocean access token. Follow instructions in the Code Ocean user guide.Read the full README ↗Model Context Protocol (MCP) server for Code Ocean.
This MCP server provides tools to search, run, and release capsules and pipelines, and manage data assets.
uv from Astral or the GitHub READMEuv python install 3.10 (or a more recent version)Each release of this Code Ocean MCP Server is tested and verified against a specific minimum version of the Code Ocean platform API. Generally, this minimum version is the latest Code Ocean version at the time of the MCP Server release. We recommend ensuring your MCP Server dependency is pinned to a version compatible with your Code Ocean deployment. For details on when the minimum Code Ocean platform version changes, see the CHANGELOG.
Here's an example VS Code MCP server configuration:
{
...
"mcp": {
"inputs": [
{
"type": "promptString",
"id": "codeocean-token",
"description": "Code Ocean API Key",
"password": true
}
],
"servers": {
"codeocean": {
"type": "stdio",
"command": "uvx",
"args": ["codeocean-mcp-server"],
"env": {
"CODEOCEAN_DOMAIN": "https://codeocean.acme.com",
"CODEOCEAN_TOKEN": "${input:codeocean-token}",
"AGENT_ID": "VS Code"
}
}
},
}
}
claude_desktop_config.json file:~/Library/Application Support/Claude/claude_desktop_config.json%APPDATA%\Claude\claude_desktop_config.json{
"mcpServers": {
"codeocean": {
"command": "uvx",
"args": ["codeocean-mcp-server"],
"env": {
"CODEOCEAN_DOMAIN": "https://codeocean.acme.com",
"CODEOCEAN_TOKEN": "<YOUR_API_KEY>",
"AGENT_ID": "Claude Desktop"
}
}
}
}
Cline stores all of its MCP settings in a JSON file called cline_mcp_settings.json. You can edit this either through the GUI (“Configure MCP Servers” in the MCP Servers pane) or by hand:
{
"mcpServers": {
"codeocean": {
"command": "uvx",
"args": ["codeocean-mcp-server"],
"env": {
"CODEOCEAN_DOMAIN": "https://codeocean.acme.com",
"CODEOCEAN_TOKEN": "<YOUR_API_KEY>",
"AGENT_ID": "Cline"
},
"alwaysAllow": [], // optional: list of tools to auto-approve
"disabled": false // ensure it’s enabled
}
}
}
Roo Code’s MCP support is configured globally across all workspaces via a JSON settings file or through its dedicated MCP Settings UI
{
"mcpServers": {
"codeocean": {
"command": "uvx",
"args": ["codeocean-mcp-server"],
"env": {
"CODEOCEAN_DOMAIN": "https://codeocean.acme.com",
"CODEOCEAN_TOKEN": "<YOUR_API_KEY>",
"AGENT_ID": "Roo Code"
}
}
}
}
Cursor stores MCP servers in a JSON file at either ~/.cursor/mcp.json (global) or {project}/.cursor/mcp.json (project-specific) .
{
"mcpServers": {
"codeocean": {
"command": "uvx",
"args": ["codeocean-mcp-server"],
"env": {
"CODEOCEAN_DOMAIN": "https://codeocean.acme.com",
"CODEOCEAN_TOKEN": "<YOUR_API_KEY>",
"AGENT_ID": "Cursor"
}
}
}
}
Windsurf (Cascade) uses mcp_config.json under ~/.codeium/windsurf/ (or via the Cascade → MCP Servers UI) .
{
"mcpServers": {
"codeocean": {
"command": "uvx",
"args": ["codeocean-mcp-server"],
"env": {
"CODEOCEAN_DOMAIN": "https://codeocean.acme.com",
"CODEOCEAN_TOKEN": "<YOUR_API_KEY>",
"AGENT_ID": "Windsurf"
}
}
}
}
By default the server runs over stdio and authenticates with the CODEOCEAN_TOKEN environment variable, as
described above. It can also serve multiple users from a single process over streamable HTTP, taking each
caller's API token from the request instead:
CODEOCEAN_DOMAIN=https://acmecorp.codeocean.com codeocean-mcp-server --transport streamable-http --host 127.0.0.1 --port 8000
Clients then pass their own token as Authorization: Bearer <YOUR_API_KEY> on every request; CODEOCEAN_TOKEN
is not used, and a request without a token is refused. The endpoint is http://<host>:<port>/mcp.
You can test the MCP server locally during development with MCP Inspector:
npx @modelcontextprotocol/inspector uv tool run codeocean-mcp-server
This will start a web server where you can:
The MCP server supports custom log formatting through the LOG_FORMAT environment variable. This allows you to control the format of log messages output by the server.
Example Format Strings: "%(asctime)s %(levelname)s [%(name)s] %(message)s".
If LOG_FORMAT is not set, the server uses FastMCP's default logging configuration.
The LOG_LEVEL environment variable sets the level the server logs at: DEBUG, INFO (the default), WARNING, ERROR or CRITICAL. Raising it drops the record the server writes for every request it serves, which is worth doing when the server runs alongside a caller that logs those requests itself.