Documentation

Quick StartFrom download to your first conversation — up and running in five minutes. Install & updateInstall ChengYoung, verify the installer, and stay up to date with in-app updates.

Local models & providersBuilt-in local inference (fully offline) or bring your own cloud provider — keys stay in the OS keychain. WorkflowsPackage instructions, parameters, and settings into reusable, shareable one-click workflows. Extensions (MCP)Give the agent new tools with extensions — built-ins work out of the box, third-party MCP servers install only with your explicit consent. SkillsTeach the agent domain know-how with skills — from simple checklists to full professional procedures, loaded on demand. SchedulesHand workflows to a timer — run automatically on your machine by cron rules, with no server in the loop. Session managementSession history, pinning and archiving, in-session actions, and automatic context management — keep long conversations under control. Context controlUse hint files, persistent instructions and prompt templates to turn project conventions, personal preferences and output rules into reusable context. PluginsExtend the agent with bundled skills, MCP servers and lifecycle hooks — installation requires explicit consent, hooks run commands on your machine, review before enabling. Codebase analysisUnderstand code structure with the built-in analysis tool — project overviews, file semantics and cross-file symbol tracking for refactoring and debugging.

Permissions & safetyTool calls are Allow / Ask / Deny — your call. Browser capabilities are authorized per domain; sensitive fields are always refused. Privacy & dataZero usage data collection; everything stays on your machine; the three network scenarios, spelled out.

FAQModel downloads, GPU, failed updates, full offline use — the eight questions we hear most. Quick tipsPractical habits for steadier, faster agent work — from plain-language prompting to managing extensions, protecting files and staying updated. Rate limits & quotasCloud model providers enforce rate limits — understand when they trigger, how to cope, and why local models are entirely unaffected.