Jump to content

Connect SuperML | Leeroopedia MCP: Equip your AI agents with best practices, code verification, and debugging knowledge. Powered by Leeroo — building Organizational Superintelligence. Contact us at founders@leeroo.com.

Implementation:OpenHands OpenHands Config Template

From Leeroopedia
Knowledge Sources
Domains Configuration, Deployment
Last Updated 2026-02-11 21:00 GMT

Overview

Reference configuration template documenting all available OpenHands settings with defaults and descriptions.

Description

The config.template.toml file serves as the authoritative reference for every configurable parameter in the OpenHands platform. It is organized into major sections: [core] (workspace, caching, iteration limits), [llm] (model selection, API keys, retry policies, safety settings), [agent] (tool toggles, browsing, Jupyter), [sandbox] (Docker settings, GPU, timeouts), [security] (confirmation mode, analyzers), [condenser] (history compression strategies), [kubernetes] (namespace, resources, PVCs), and [mcp] (server configurations).

Usage

Use this template when deploying or configuring an OpenHands instance. Copy to config.toml and customize values for your environment. This is the starting point for all deployment configurations.

Code Reference

Source Location

Configuration Sections

[core]
workspace_base = "./workspace"
cache_dir = "/tmp/cache"
max_iterations = 100

[llm]
model = "anthropic/claude-sonnet-4-20250514"
api_key = ""
temperature = 0.0

[agent]
enable_browsing = false
enable_jupyter = false

[sandbox]
base_container_image = "nikolaik/python-nodejs:python3.12-nodejs22"
timeout = 120

[security]
confirmation_mode = false

[condenser]
type = "noop"

[kubernetes]
namespace = "openhands"

[mcp]
# Server configurations for SSE, SHTTP, stdio

Import

from openhands.core.config import OpenHandsConfig
# The template is loaded and parsed by OpenHandsConfig

I/O Contract

Inputs

Section Key Parameters Type Description
[core] workspace_base, max_iterations, max_budget_per_task string, int, float Core runtime limits and paths
[llm] model, api_key, temperature, top_p string, string, float, float LLM provider configuration
[agent] enable_browsing, enable_jupyter bool, bool Agent capability toggles
[sandbox] base_container_image, timeout string, int Sandbox environment settings
[condenser] type string History compression strategy (noop, observation_masking, recent, llm, amortized)

Outputs

Name Type Description
OpenHandsConfig object Fully parsed configuration object used throughout the application

Usage Examples

Minimal Configuration

[core]
workspace_base = "./workspace"

[llm]
model = "anthropic/claude-sonnet-4-20250514"
api_key = "sk-ant-..."

Production Configuration

[core]
workspace_base = "/data/workspace"
max_iterations = 200
max_budget_per_task = 5.0

[llm]
model = "anthropic/claude-sonnet-4-20250514"
api_key = "sk-ant-..."
num_retries = 5
retry_multiplier = 2.0

[sandbox]
base_container_image = "custom-image:latest"
timeout = 300

[security]
confirmation_mode = true

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment