Heartbeats
Heartbeats are a mechanism that enables Letta agents to chain multiple tool calls together in a single execution loop. The term “heartbeat” was coined in the MemGPT paper, and since the Letta codebase evolved from the original MemGPT codebase (same authors), heartbeats remain a core part of the default agent loop.
How heartbeats work
Every tool in Letta automatically receives an additional parameter called request_heartbeat, which defaults to false. When an agent sets this parameter to true, it signals to the Letta server that it wants to continue executing after the current tool call completes.
Technical implementation
When the Letta server detects that request_heartbeat=true, it:
- Completes the current tool execution
- Restarts the agent loop with a system message acknowledging the heartbeat request
- Allows the agent to continue with an additional tool calls
This enables agents to perform complex, multi-step operations without requiring explicit user intervention between steps.
Automatic heartbeats on failure
If a tool call fails at runtime, Letta automatically generates a heartbeat. This gives the agent an opportunity to handle the error and potentially retry the operation with different parameters or take alternative actions.
Viewing heartbeats in the ADE
In the Agent Development Environment (ADE), heartbeat requests are visible for all agent messages.
When a tool is called with request_heartbeat=true, you’ll see a heartbeat indicator next to the tool call, making it easy to track when an agent is proactively chaining operations together.
Learn more
To read more about the concept of heartbeats and their origins, refer to the original MemGPT research paper.