176 lines
2.9 KiB
Markdown
176 lines
2.9 KiB
Markdown
---
|
|
sidebar_position: 3
|
|
---
|
|
|
|
# MCP API
|
|
|
|
Model Context Protocol (MCP) tools exposed by the Memora MCP server.
|
|
|
|
## Available Tools
|
|
|
|
### memora_search
|
|
|
|
Search memories for an agent.
|
|
|
|
**Parameters:**
|
|
|
|
| Parameter | Type | Required | Description |
|
|
|-----------|------|----------|-------------|
|
|
| `query` | string | yes | Search query |
|
|
| `agent_id` | string | no | Agent ID (uses default if not specified) |
|
|
| `top_k` | integer | no | Number of results (default: 10) |
|
|
|
|
**Example:**
|
|
|
|
```json
|
|
{
|
|
"name": "memora_search",
|
|
"arguments": {
|
|
"query": "What does Alice do for work?",
|
|
"top_k": 5
|
|
}
|
|
}
|
|
```
|
|
|
|
**Response:**
|
|
|
|
```json
|
|
{
|
|
"results": [
|
|
{
|
|
"text": "Alice works at Google as a software engineer",
|
|
"weight": 0.95,
|
|
"fact_type": "world"
|
|
}
|
|
]
|
|
}
|
|
```
|
|
|
|
---
|
|
|
|
### memora_think
|
|
|
|
Generate a personality-aware response using retrieved memories.
|
|
|
|
**Parameters:**
|
|
|
|
| Parameter | Type | Required | Description |
|
|
|-----------|------|----------|-------------|
|
|
| `query` | string | yes | Question or prompt |
|
|
| `agent_id` | string | no | Agent ID (uses default if not specified) |
|
|
| `thinking_budget` | integer | no | Tokens for reasoning (default: 100) |
|
|
|
|
**Example:**
|
|
|
|
```json
|
|
{
|
|
"name": "memora_think",
|
|
"arguments": {
|
|
"query": "What should I recommend to Alice?"
|
|
}
|
|
}
|
|
```
|
|
|
|
**Response:**
|
|
|
|
```json
|
|
{
|
|
"text": "Based on Alice's interest in machine learning and her work at Google, I would recommend...",
|
|
"based_on": [
|
|
{"text": "Alice works at Google", "weight": 0.95}
|
|
],
|
|
"new_opinions": []
|
|
}
|
|
```
|
|
|
|
---
|
|
|
|
### memora_store
|
|
|
|
Store a new memory.
|
|
|
|
**Parameters:**
|
|
|
|
| Parameter | Type | Required | Description |
|
|
|-----------|------|----------|-------------|
|
|
| `content` | string | yes | Memory content to store |
|
|
| `agent_id` | string | no | Agent ID (uses default if not specified) |
|
|
| `context` | string | no | Context or topic of the memory |
|
|
|
|
**Example:**
|
|
|
|
```json
|
|
{
|
|
"name": "memora_store",
|
|
"arguments": {
|
|
"content": "User prefers Python for data analysis",
|
|
"context": "programming discussion"
|
|
}
|
|
}
|
|
```
|
|
|
|
**Response:**
|
|
|
|
```json
|
|
{
|
|
"success": true,
|
|
"message": "Memory stored successfully"
|
|
}
|
|
```
|
|
|
|
---
|
|
|
|
### memora_agents
|
|
|
|
List all available agents.
|
|
|
|
**Parameters:** None
|
|
|
|
**Example:**
|
|
|
|
```json
|
|
{
|
|
"name": "memora_agents",
|
|
"arguments": {}
|
|
}
|
|
```
|
|
|
|
**Response:**
|
|
|
|
```json
|
|
{
|
|
"agents": [
|
|
{"agent_id": "default"},
|
|
{"agent_id": "assistant"},
|
|
{"agent_id": "researcher"}
|
|
]
|
|
}
|
|
```
|
|
|
|
---
|
|
|
|
## Environment Variables
|
|
|
|
| Variable | Description | Default |
|
|
|----------|-------------|---------|
|
|
| `MEMORA_API_URL` | Memora API URL | `http://localhost:8080` |
|
|
| `MEMORA_AGENT_ID` | Default agent ID | Required |
|
|
|
|
## Error Responses
|
|
|
|
MCP tools return errors in the standard MCP error format:
|
|
|
|
```json
|
|
{
|
|
"error": {
|
|
"code": "NOT_FOUND",
|
|
"message": "Agent 'unknown-agent' not found"
|
|
}
|
|
}
|
|
```
|
|
|
|
| Code | Description |
|
|
|------|-------------|
|
|
| `INVALID_PARAMS` | Missing or invalid parameters |
|
|
| `NOT_FOUND` | Agent or resource not found |
|
|
| `INTERNAL_ERROR` | Server error |
|