Files
OpenHands/openhands
openhands 1231bd3de1 refactor: derive all model lists from openhands-sdk (single source of truth)
Replace the hardcoded OPENHANDS_MODELS list and provider-assignment
tables in llm.py with imports from the openhands-sdk package:

  from openhands.sdk.llm.utils.verified_models import (
      VERIFIED_MODELS, VERIFIED_OPENHANDS_MODELS,
      VERIFIED_OPENAI_MODELS, VERIFIED_ANTHROPIC_MODELS,
      VERIFIED_MISTRAL_MODELS,
  )

This means:
- OPENHANDS_MODELS is now built dynamically from the SDK's
  VERIFIED_OPENHANDS_MODELS
- VERIFIED_PROVIDERS is derived from VERIFIED_MODELS.keys()
- _BARE_OPENAI_MODELS, _BARE_ANTHROPIC_MODELS, _BARE_MISTRAL_MODELS
  are sets built from the SDK's per-provider lists
- The SDK is the single source of truth for both V0/V1 conversations

To add or remove a verified model, update the SDK
(openhands-sdk verified_models.py) — no change needed in this repo.

Co-authored-by: openhands <openhands@all-hands.dev>
2026-03-16 18:30:34 +00:00
..
2026-02-25 05:11:13 -07:00
2025-10-14 02:16:44 +00:00
2024-08-26 01:31:37 -06:00
2025-10-14 02:16:44 +00:00

OpenHands Architecture

This directory contains the core components of OpenHands.

For an overview of the system architecture, see the architecture documentation (v0 backend architecture).

Classes

The key classes in OpenHands are:

  • LLM: brokers all interactions with large language models. Works with any underlying completion model, thanks to LiteLLM.
  • Agent: responsible for looking at the current State, and producing an Action that moves one step closer toward the end-goal.
  • AgentController: initializes the Agent, manages State, and drive the main loop that pushes the Agent forward, step by step
  • State: represents the current state of the Agent's task. Includes things like the current step, a history of recent events, the Agent's long-term plan, etc
  • EventStream: a central hub for Events, where any component can publish Events, or listen for Events published by other components
    • Event: an Action or Observeration
      • Action: represents a request to e.g. edit a file, run a command, or send a message
      • Observation: represents information collected from the environment, e.g. file contents or command output
  • Runtime: responsible for performing Actions, and sending back Observations
    • Sandbox: the part of the runtime responsible for running commands, e.g. inside of Docker
  • Server: brokers OpenHands sessions over HTTP, e.g. to drive the frontend
    • Session: holds a single EventStream, a single AgentController, and a single Runtime. Generally represents a single task (but potentially including several user prompts)
    • ConversationManager: keeps a list of active sessions, and ensures requests are routed to the correct Session

Control Flow

Here's the basic loop (in pseudocode) that drives agents.

while True:
  prompt = agent.generate_prompt(state)
  response = llm.completion(prompt)
  action = agent.parse_response(response)
  observation = runtime.run(action)
  state = state.update(action, observation)

In reality, most of this is achieved through message passing, via the EventStream. The EventStream serves as the backbone for all communication in OpenHands.

flowchart LR
  Agent--Actions-->AgentController
  AgentController--State-->Agent
  AgentController--Actions-->EventStream
  EventStream--Observations-->AgentController
  Runtime--Observations-->EventStream
  EventStream--Actions-->Runtime
  Frontend--Actions-->EventStream

Runtime

Please refer to the documentation to learn more about Runtime.