https://github.com/loopwork-ai/emcee
Connect agents to APIs 🫳🎤💥
https://github.com/loopwork-ai/emcee
agents modelcontextprotocol openapi
Last synced: 2 months ago
JSON representation
Connect agents to APIs 🫳🎤💥
- Host: GitHub
- URL: https://github.com/loopwork-ai/emcee
- Owner: loopwork-ai
- License: apache-2.0
- Created: 2025-01-16T16:20:26.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2025-03-29T11:44:13.000Z (3 months ago)
- Last Synced: 2025-04-02T04:05:37.317Z (3 months ago)
- Topics: agents, modelcontextprotocol, openapi
- Language: Go
- Homepage: https://emcee.sh
- Size: 149 KB
- Stars: 188
- Watchers: 2
- Forks: 11
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- trackawesomelist - loopwork-ai/emcee (⭐118)
README
# emcee
**emcee** is a tool that provides a [Model Context Protocol (MCP)][mcp] server
for any web application with an [OpenAPI][openapi] specification.
You can use emcee to connect [Claude Desktop][claude] and [other apps][mcp-clients]
to external tools and data services,
similar to [ChatGPT plugins][chatgpt-plugins].
## Quickstart
If you're on macOS and have [Homebrew][homebrew] installed,
you can get up-and-running quickly.
```bash
# Install emcee
brew install loopwork-ai/tap/emcee
```
Make sure you have [Claude Desktop](https://claude.ai/download) installed.
To configure Claude Desktop for use with emcee:
1. Open Claude Desktop Settings (⌘,)
2. Select the "Developer" section in the sidebar
3. Click "Edit Config" to open the configuration file
The configuration file should be located in the Application Support directory.
You can also open it directly in VSCode using:
```console
code ~/Library/Application\ Support/Claude/claude_desktop_config.json
```
Add the following configuration to add the weather.gov MCP server:
```json
{
"mcpServers": {
"weather": {
"command": "emcee",
"args": [
"https://api.weather.gov/openapi.json"
]
}
}
}
```
After saving the file, quit and re-open Claude.
You should now see 🔨57 in the bottom right corner of your chat box.
Click on that to see a list of all the tools made available to Claude through MCP.
Start a new chat and ask it about the weather where you are.
> What's the weather in Portland, OR?
Claude will consult the tools made available to it through MCP
and request to use one if deemed to be suitable for answering your question.
You can review this request and either approve or deny it.
If you allow, Claude will communicate with the MCP
and use the result to inform its response.
---
> [!TIP]
> Building agents? Want to deploy remote MCP servers?
> Reach out to us at [email protected]
---
## Why use emcee?
MCP provides a standardized way to connect AI models to tools and data sources.
It's still early days, but there are already a variety of [available servers][mcp-servers]
for connecting to browsers, developer tools, and other systems.
We think emcee is a convenient way to connect to services
that don't have an existing MCP server implementation —
_especially for services you're building yourself_.
Got a web app with an OpenAPI spec?
You might be surprised how far you can get
without a dashboard or client library.
## Installation
### Installer Script
Use the [installer script][installer] to download and install a
[pre-built release][releases] of emcee for your platform
(Linux x86-64/i386/arm64 and macOS Intel/Apple Silicon).
```console
# fish
sh (curl -fsSL https://get.emcee.sh | psub)
# bash, zsh
sh <(curl -fsSL https://get.emcee.sh)
```
### Homebrew
Install emcee using [Homebrew][homebrew] from [Loopwork's tap][homebrew-tap].
```console
brew install loopwork-ai/tap/emcee
```
### Docker
Prebuilt [Docker images][docker-images] with emcee are available.
```console
docker run -it ghcr.io/loopwork-ai/emcee
```
### Build From Source
Requires [go 1.24][golang] or later.
```console
git clone https://github.com/loopwork-ai/emcee.git
cd emcee
go build -o emcee cmd/emcee/main.go
```
Once built, you can run in place (`./emcee`)
or move it somewhere in your `PATH`, like `/usr/local/bin`.
## Usage
```console
Usage:
emcee [spec-path-or-url] [flags]
Flags:
--basic-auth string Basic auth value (either user:pass or base64 encoded, will be prefixed with 'Basic ')
--bearer-auth string Bearer token value (will be prefixed with 'Bearer ')
-h, --help help for emcee
--raw-auth string Raw value for Authorization header
--retries int Maximum number of retries for failed requests (default 3)
-r, --rps int Maximum requests per second (0 for no limit)
-s, --silent Disable all logging
--timeout duration HTTP request timeout (default 1m0s)
-v, --verbose Enable debug level logging to stderr
--version version for emcee
```
emcee implements [Standard Input/Output (stdio)](https://modelcontextprotocol.io/docs/concepts/transports#standard-input-output-stdio) transport for MCP,
which uses [JSON-RPC 2.0](https://www.jsonrpc.org/) as its wire format.
When you run emcee from the command-line,
it starts a program that listens on stdin,
outputs to stdout,
and logs to stderr.
### Authentication
For APIs that require authentication,
emcee supports several authentication methods:
| Authentication Type | Example Usage | Resulting Header |
|------------------------|---------------|----------------------------|
| **Bearer Token** | `--bearer-auth="abc123"` | `Authorization: Bearer abc123` |
| **Basic Auth** | `--basic-auth="user:pass"` | `Authorization: Basic dXNlcjpwYXNz` |
| **Raw Value** | `--raw-auth="Custom xyz789"` | `Authorization: Custom xyz789` |
These authentication values can be provided directly
or as [1Password secret references][secret-reference-syntax].
When using 1Password references:
- Use the format `op://vault/item/field`
(e.g. `--bearer-auth="op://Shared/X/credential"`)
- Ensure the 1Password CLI ([op][op]) is installed and available in your `PATH`
- Sign in to 1Password before running emcee or launching Claude Desktop
```console
# Install op
brew install 1password-cli
# Sign in 1Password CLI
op signin
```
```json
{
"mcpServers": {
"twitter": {
"command": "emcee",
"args": [
"--bearer-auth=op://shared/x/credential",
"https://api.twitter.com/2/openapi.json"
]
}
}
}
```
> [!IMPORTANT]
> emcee doesn't use auth credentials when downloading
> OpenAPI specifications from URLs provided as command arguments.
> If your OpenAPI specification requires authentication to access,
> first download it to a local file using your preferred HTTP client,
> then provide the local file path to emcee.
### Transforming OpenAPI Specifications
You can transform OpenAPI specifications before passing them to emcee using standard Unix utilities. This is useful for:
- Selecting specific endpoints to expose as tools
with [jq][jq] or [yq][yq]
- Modifying descriptions or parameters
with [OpenAPI Overlays][openapi-overlays]
- Combining multiple specifications
with [Redocly][redocly-cli]
For example,
you can use `jq` to include only the `point` tool from `weather.gov`.
```console
cat path/to/openapi.json | \
jq 'if .paths then .paths |= with_entries(select(.key == "/points/{point}")) else . end' | \
emcee
```
### JSON-RPC
You can interact directly with the provided MCP server
by sending JSON-RPC requests.
> [!NOTE]
> emcee provides only MCP tool capabilities.
> Other features like resources, prompts, and sampling aren't yet supported.
#### List Tools
Request
```json
{"jsonrpc": "2.0", "method": "tools/list", "params": {}, "id": 1}
```
Response
```jsonc
{
"jsonrpc":"2.0",
"result": {
"tools": [
// ...
{
"name": "tafs",
"description": "Returns Terminal Aerodrome Forecasts for the specified airport station.",
"inputSchema": {
"type": "object",
"properties": {
"stationId": {
"description": "Observation station ID",
"type": "string"
}
},
"required": ["stationId"]
}
},
// ...
]
},
"id": 1
}
```
#### Call Tool
Request
```json
{"jsonrpc": "2.0", "method": "tools/call", "params": { "name": "taf", "arguments": { "stationId": "KPDX" }}, "id": 1}
```
Response
```jsonc
{
"jsonrpc":"2.0",
"content": [
{
"type": "text",
"text": /* Weather forecast in GeoJSON format */,
"annotations": {
"audience": ["assistant"]
}
}
]
"id": 1
}
```
## Debugging
The [MCP Inspector][mcp-inspector] is a tool for testing and debugging MCP servers.
If Claude and/or emcee aren't working as expected,
the inspector can help you understand what's happening.
```console
npx @modelcontextprotocol/inspector emcee https://api.weather.gov/openapi.json
# 🔍 MCP Inspector is up and running at http://localhost:5173 🚀
```
```console
open http://localhost:5173
```
## License
emcee is licensed under the Apache License, Version 2.0.
[chatgpt-plugins]: https://openai.com/index/chatgpt-plugins/
[claude]: https://claude.ai/download
[docker-images]: https://github.com/loopwork-ai/emcee/pkgs/container/emcee
[golang]: https://go.dev
[homebrew]: https://brew.sh
[homebrew-tap]: https://github.com/loopwork-ai/homebrew-tap
[installer]: https://github.com/loopwork-ai/emcee/blob/main/tools/install.sh
[jq]: https://github.com/jqlang/jq
[mcp]: https://modelcontextprotocol.io/
[mcp-clients]: https://modelcontextprotocol.info/docs/clients/
[mcp-inspector]: https://github.com/modelcontextprotocol/inspector
[mcp-servers]: https://modelcontextprotocol.io/examples
[op]: https://developer.1password.com/docs/cli/get-started/
[openapi]: https://openapi.org
[openapi-overlays]: https://www.openapis.org/blog/2024/10/22/announcing-overlay-specification
[redocly-cli]: https://redocly.com/docs/cli/commands
[releases]: https://github.com/loopwork-ai/emcee/releases
[secret-reference-syntax]: https://developer.1password.com/docs/cli/secret-reference-syntax/
[yq]: https://github.com/mikefarah/yq