Atoms MCP Server
# Atoms MCP Server
MCP server for the [Atoms](https://atoms.smallest.ai) voice AI platform. Manage agents, debug calls, view analytics — directly from your IDE.
## Quick start
### Option A: One-line installer (no dependencies)
```bash
curl -fsSL https://raw.githubusercontent.com/aryasmol/atoms-mcp-server/main/install.sh | bash
```
Downloads a standalone binary, prompts for your API key, and configures Cursor + Claude Desktop automatically.
### Option B: npm
Requires Node.js 18+. Add this to your MCP config (`~/.cursor/mcp.json` for Cursor, `claude_desktop_config.json` for Claude Desktop):
```json
{
"mcpServers": {
"atoms": {
"command": "npx",
"args": ["-y", "@developer-smallestai/atoms-mcp-server"],
"env": {
"ATOMS_API_KEY": "sk_your_key_here"
}
}
}
}
```
### Option C: One prompt
Open a chat in Cursor or Claude Desktop and type:
```
Set up the Atoms MCP server for me.
The npm package is @developer-smallestai/atoms-mcp-server.
My API key is: sk_paste_your_key_here
```
### Verify
Reload your editor, then type: **"List all my agents"**
---
## Available tools
### Read
| Tool | Description |
|---|---|
| `get_call_logs` | Query call logs with filters for status, type, date range, agent name, or phone number |
| `debug_call` | Deep-dive into a single call — full transcript, errors, timing, cost breakdown, post-call analytics |
| `get_agents` | List agents with their configuration, voice, LLM model, and call stats |
| `get_usage_stats` | Usage statistics — total calls, duration, costs, status breakdown |
| `get_campaigns` | List outbound calling campaigns with status and progress |
| `get_phone_numbers` | List phone numbers owned by your organization |
### Write
| Tool | Description |
|---|---|
| `create_agent` | Create a new AI voice agent |
| `update_agent_prompt` | Update an agent's system prompt / instructions |
| `update_agent_config` | Update agent settings — name, language, voice, first message, etc. |
| `delete_agent` | Archive (soft-delete) an agent |
### Act
| Tool | Description |
|---|---|
| `make_call` | Initiate an outbound phone call using a specific agent |
### Resources
| Resource | URI | Description |
|---|---|---|
| Platform Overview | `atoms://docs/platform-overview` | Key concepts, call types, statuses, and cost breakdown |
---
## Environment variables
| Variable | Required | Default | Description |
|---|---|---|---|
| `ATOMS_API_KEY` | Yes | — | Your Atoms API key |
| `ATOMS_API_URL` | No | `https://atoms-api.smallest.ai/api/v1` | Override the API base URL |
## Development
```bash
npm install
npm run dev # run with tsx
npm run build # bundle to dist/
```
## License
MIT
TDQS
Scored across 11 tools
Each tool maps to a distinct resource or action—agent CRUD, call logs, aggregate usage, single-call debugging, campaigns, phone numbers, and outbound calls. The call-related tools are clearly separated by scope (list vs. aggregate vs. single-call deep-dive), so an agent can reliably select the right one.
All names follow a consistent snake_case verb_noun pattern: get_* for read operations, create_/update_/delete_ for agent mutations, and make_call/debug_call for actions. There are no mixed conventions or vague verbs.
11 tools is well-scoped for an AI-calling platform: enough to cover meaningful agent and call workflows without redundancy. The set feels deliberately chosen rather than bloated or thin.
Agent lifecycle is fully covered with create, list, config update, prompt update, and delete, and call workflows are covered via make_call, get_call_logs, debug_call, and get_usage_stats. Campaigns and phone numbers are read-only, and there is no singular get_agent by ID, but these are minor gaps that agents can typically work around.