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

# A2A Integration

> Use the standard A2A protocol to communicate with MemoryLake agents

MemoryLake implements the [A2A (Agent-to-Agent) protocol](https://a2a-protocol.org/) by Google. If you are already familiar with A2A, you can use any A2A-compatible client to interact with MemoryLake agents.

Refer to the [A2A specification](https://a2a-protocol.org/latest/specification/) for the standard request and response schemas. Official SDKs are available at [github.com/a2aproject](https://github.com/a2aproject).

<Note>
  **Permission required:** `workspace:a2a`
</Note>

## Base URL

```
https://app.memorylake.ai/openapi/memorylake/api/v3/workspaces/{workspaceId}/agents/{agentId}/a2a
```

Append standard A2A paths directly — for example `POST {baseUrl}/message:send` or `GET {baseUrl}/tasks/{taskId}`.

## MemoryLake Extension

MemoryLake extends the standard A2A protocol through the `metadata.memorylake` namespace. Pass this object in the request metadata of `message:send` or `message:stream` to control how the agent interacts with MemoryLake's memory layer.

This is the key difference from a vanilla A2A server — these fields connect the agent to workspaces, projects, and actors in MemoryLake.

<ParamField body="metadata.memorylake" type="object" required>
  MemoryLake-specific configuration. Without this, the agent runs without any memory context.

  <Expandable title="Fields" defaultOpen>
    <ParamField body="actorId" type="string">
      Actor identifier representing the end user. Facts extracted from the conversation are associated with this actor, enabling per-user memory.
    </ParamField>

    <ParamField body="readWriteProjectId" type="string">
      Project the agent can both read from and write to. Extracted facts and conversation history are persisted to this project.
    </ParamField>

    <ParamField body="readOnlyProjectIds" type="array">
      Additional projects the agent can search (documents and facts) but not modify. Use this to give the agent access to shared knowledge bases.
    </ParamField>

    <ParamField body="subagentMapping" type="object">
      Maps subagent names (defined in the agent configuration) to concrete agent IDs. Allows dynamic routing of subagent calls at request time.
    </ParamField>

    <ParamField body="overrides" type="object">
      Runtime overrides for the agent's configuration.

      <Expandable title="Override fields">
        <ParamField body="model" type="string">Override the agent's model (e.g. `gpt-4o`, `claude-sonnet-4-20250514`)</ParamField>
        <ParamField body="systemPromptOverride" type="string">Replace the agent's system prompt for this request</ParamField>
        <ParamField body="modelParams" type="object">Override model parameters (e.g. `temperature`, `max_tokens`)</ParamField>

        <ParamField body="reasoning" type="object">
          Reasoning configuration

          <Expandable title="Reasoning fields">
            <ParamField body="type" type="string">Reasoning type</ParamField>
            <ParamField body="budgetTokens" type="integer">Token budget for reasoning</ParamField>
          </Expandable>
        </ParamField>

        <ParamField body="capabilitiesOverride" type="array">Override the agent's capabilities list</ParamField>
        <ParamField body="maxTurns" type="integer">Override the maximum number of turns</ParamField>
      </Expandable>
    </ParamField>
  </Expandable>
</ParamField>

## Example

<RequestExample>
  ```bash Send message with memory context theme={null}
  curl -X POST 'https://app.memorylake.ai/openapi/memorylake/api/v3/workspaces/ws_abc123/agents/agt_r8k2m1n3/a2a/message:send' \
    -H 'Authorization: Bearer sk_xxxxxx' \
    -H 'Content-Type: application/json' \
    -d '{
      "message": {
        "role": "user",
        "content": [
          { "text": "How do I configure two-factor authentication?" }
        ],
        "contextId": "ctx_m4n7p2q9"
      },
      "configuration": {
        "blocking": true,
        "historyLength": 10
      },
      "metadata": {
        "memorylake": {
          "actorId": "act_user_jane",
          "readWriteProjectId": "proj_def456",
          "readOnlyProjectIds": ["proj_shared_kb"],
          "overrides": {
            "model": "gpt-4o",
            "maxTurns": 5
          }
        }
      }
    }'
  ```

  ```bash Stream with memory context (SSE) theme={null}
  curl -X POST 'https://app.memorylake.ai/openapi/memorylake/api/v3/workspaces/ws_abc123/agents/agt_r8k2m1n3/a2a/message:stream' \
    -H 'Authorization: Bearer sk_xxxxxx' \
    -H 'Content-Type: application/json' \
    -H 'Accept: text/event-stream' \
    -d '{
      "message": {
        "role": "user",
        "content": [
          { "text": "Summarize our last meeting notes" }
        ]
      },
      "metadata": {
        "memorylake": {
          "actorId": "act_user_jane",
          "readOnlyProjectIds": ["proj_meetings"]
        }
      }
    }'
  ```
</RequestExample>