Execution Modes

Execution Modes

Async-capable tools support:

• auto
• sync
• async

auto

The server chooses the best mode.

Typical behavior:

• small requests stay sync
• larger, batched, or heavier requests may switch to async

sync

The tool waits for the direct response and returns data immediately.

Use this when:

• the request is small
• you want a direct answer
• deterministic immediate output is preferred

async

The tool submits a request and returns async metadata.

Use this when:

• the request is large
• you expect longer processing
• you want better reliability for heavier endpoints

Async Workflow

1. Call a tool with execution_mode: "async" or execution_mode: "auto".
2. Receive async metadata with request ID.
3. Poll status with requests_get.
4. Use requests_list to inspect running or finished requests.
5. Use requests_delete if you want to stop a request.

Example:

{
  "query": ["outscraper.com"],
  "execution_mode": "async"
}