> ## Documentation Index
> Fetch the complete documentation index at: https://docs.minns.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Search by context

> Find memories relevant to the current task by matching goals and environment against past episodes.

This is the most powerful memory retrieval method. It answers: *"What did we do last time we were in this exact situation?"*

## Request body

<ParamField body="context" type="EventContext" required>
  Context to match against. See [Types](/api-reference/types#eventcontext-structure).
</ParamField>

<ParamField body="limit" type="integer" default="10">
  Maximum number of results.
</ParamField>

<ParamField body="min_similarity" type="number" default="0.6">
  Minimum similarity threshold (`0.0`–`1.0`). Set to `null` to use the default.
</ParamField>

<ParamField body="agent_id" type="integer">
  Filter by agent. Set to `null` to search across all agents.
</ParamField>

<ParamField body="session_id" type="integer">
  Filter by session. Set to `null` to search across all sessions.
</ParamField>

### Example request

```json theme={null}
{
  "context": {
    "active_goals": [
      {
        "id": 101,
        "description": "book_movie",
        "priority": 0.9,
        "progress": 0.5,
        "deadline": null,
        "subgoals": []
      }
    ],
    "environment": {
      "variables": { "user_id": "user_99" },
      "spatial": null,
      "temporal": {
        "time_of_day": null,
        "deadlines": [],
        "patterns": []
      }
    },
    "resources": {
      "computational": {
        "cpu_percent": 50.0,
        "memory_bytes": 1024000,
        "storage_bytes": 1024000000,
        "network_bandwidth": 1000
      },
      "external": {}
    },
    "fingerprint": 0,
    "embeddings": null
  },
  "limit": 5,
  "min_similarity": 0.8,
  "agent_id": null,
  "session_id": null
}
```

## Response

Returns an array of `MemoryResponse` objects (same schema as [Get agent memories](/api-reference/memory/get-agent-memories)).

<Tip>
  The system uses a **context fingerprint** (a hash of the context) to find near-identical situations instantly. Always query by context and goal — this ensures your agent doesn't get confused by history from a different task.
</Tip>