chatpharo

Configuration

ChatPharo configuration is managed by ChatPharoSettings and edited via Spec presenters.

Persistence locations

• Logs (when enabled): FileLocator imageDirectory / 'chatpharo' / 'log-chatpharo.txt'
• Settings are saved as the default configuration for the image.

Safety Advisor defaults

ChatPharo includes an Advisor that can apply recommended settings such as:

• agent selection
• enabling tools and skills
• enabling caching + feedback buttons
• enabling sandbox restrictions with a timeout
• setting a safe maximumIterations

Common settings overview

Setting	What it affects
Agent + model	Which backend answers prompts
maximumIterations	Bounds tool-call loops
Tools enabled list	Which built-in tools the model can call
Custom tools	User-defined tools exposed to the model
Skills enabled list	Which skills are always injected
automaticSkillsEnabled	Auto-inject skills via keyword detection
skillToolsEnabled	Allow the model to request skill context
Memory toggles	Retention, item limit, auto-summarize
Multivers	Whether prompt chaining is active
MCP	Whether MCP servers are connected and tools exposed
Sandbox	Restrictions/timeout used by tool execution
Logging	Whether events are written to disk

Practical guidance

• Prefer local agents (Ollama/LM Studio) if you need offline behavior.
• Keep tools enabled unless you are diagnosing problems.
• Keep maximumIterations modest (2–5) for predictable tool behavior.

Related docs

• Agents: agents.md
• Tools: tools.md
• Memory: memory.md
• Sandbox: sandbox.md

This site is open source. Improve this page.