# Empryo docs

> Documentation index for the Empryo AI coding agent. Every URL also serves raw markdown at the same path + `.md`. Whole site: https://empryo.com/llms.txt · full corpus: https://empryo.com/llms-full.txt

- [Tabs](https://empryo.com/docs/agents/cross-tab.md): Run up to 10 independent workspaces per project with tab-aware file claims, shared Genome and LSP, and git hard-blocked mid-dispatch so no partial edits.
- [Parallel agents](https://empryo.com/docs/agents/dispatch.md): Forge dispatches Spark and Ember sub-agents in parallel for multi-file work, sharing an I/O cache and findings while a model mix cuts cost.
- [Docs for agents](https://empryo.com/docs/agents/docs-mcp.md): Copy any docs page or section as Markdown or plain text, connect an agent to the read-only documentation MCP, and use the empryo skill that ships installed.
- [The reviewer](https://empryo.com/docs/agents/review.md): A fresh-context agent reads what changed and returns PASS, FAIL or PARTIAL. Run it on demand, every goal-loop round, or as a CI exit code.
- [Steering](https://empryo.com/docs/agents/steering.md): Redirect the agent mid-flight by typing while it works. Up to 5 messages queue and inject at the next step, with no abort or restart.
- [How it works](https://empryo.com/docs/concepts/architecture.md): How Empryo works: a live ranked code graph, an agent loop with parallel sub-agents, and embedded Neovim with layered LSP, AST, and tree-sitter intelligence.
- [Compound tools](https://empryo.com/docs/concepts/compound-tools.md): One-call tools for rename, move, refactor, test, and commit. LSP-backed and verified: renaming a class across 8 files takes 3 steps instead of 19.
- [Cost tracking](https://empryo.com/docs/concepts/cost-tracking.md): Live USD cost tracking per model and per sub-agent, with built-in pricing for every major provider, cache-aware billing, and a /context breakdown dashboard.
- [The Genome](https://empryo.com/docs/concepts/genome.md): How Empryo scans your repo into a ranked dependency graph (PageRank, git co-change, 30+ languages) and injects the map into every prompt.
- [Code intelligence](https://empryo.com/docs/concepts/intelligence.md): How the agent navigates code by symbol using LSP, ts-morph, tree-sitter, and regex fallbacks, plus diagnostics tracked after every edit.
- [The Marionette](https://empryo.com/docs/concepts/marionette.md): A prompt-compiler pre-pass: a cheap model walks the code graph before the main agent moves and attaches a map — files:lines, wiring, memories — to your prompt.
- [Model events](https://empryo.com/docs/concepts/model-events.md): An opt-in debug view recording every LLM call's latency, tokens, and errors so you can spot the bottleneck when routing across providers.
- [Context compaction](https://empryo.com/docs/context/compaction.md): How Empryo compacts long conversations at 70% of the context window using structured-state (V2) or LLM-summary (V1) strategies, configurably.
- [Modes & the prompt](https://empryo.com/docs/context/prompt-system.md): Switch agent modes with Ctrl+D, auto-load EMPRYO.md and other tools' instruction files, and get per-model-family prompt tuning automatically.
- [Thinking & retries](https://empryo.com/docs/context/provider-options.md): Configure thinking modes and budgets, Anthropic context management, graceful option degradation, and rate-limit retry backoff per model.
- [FAQ](https://empryo.com/docs/faq.md): Answers on Windows support, API keys vs local LLMs, run costs, how Empryo edits code via AST and LSP, privacy, and 30+ language coverage.
- [Empryo](https://empryo.com/docs/index.md): Meet Empryo, an AI coding agent for terminal, desktop, and headless workflows. Explore your codebase, make focused changes, and verify results.
- [Installation](https://empryo.com/docs/installation.md): Install Empryo on macOS, Linux, or Windows x64 using the official CLI installer or desktop download, then check your installation and connect a model.
- [GitHub Copilot](https://empryo.com/docs/providers/copilot.md): Use your paid GitHub Copilot subscription in Empryo: sign in with a device code, watch the monthly allowance, run Claude, GPT and Gemini at no extra cost.
- [Custom providers](https://empryo.com/docs/providers/custom.md): Connect any OpenAI-compatible API to Empryo with one config block: set baseURL, key env var, models, dynamic model lists, and reasoning options.
- [Providers](https://empryo.com/docs/providers/overview.md): Configure Empryo's 23 built-in LLM providers across cloud, gateways, subscriptions, local, and enterprise, plus any OpenAI-compatible custom API.
- [Proxy](https://empryo.com/docs/providers/proxy.md): Route Claude, GPT, Gemini and Grok through the CLIProxyAPI add-on, and read the plan windows it reports: per-model caps and reset countdowns.
- [Subscriptions](https://empryo.com/docs/providers/subscriptions.md): Sign in to Claude, ChatGPT, Cursor, Devin, Gemini, Kimi and more from Empryo with no add-on. One /login picker, credentials in your keychain.
- [Quickstart](https://empryo.com/docs/quickstart.md): Connect a supported account, API key, or local model to Empryo, open a trusted project, and complete your first useful task with a reviewable result.
- [Run from CI or scripts](https://empryo.com/docs/recipes/headless.md): Run Empryo without the TUI for scripts and CI: pipe a prompt, stream JSON or JSONL events, resume sessions, and cap runs with max-steps and timeout flags.
- [Model fallback chains](https://empryo.com/docs/recipes/model-fallback.md): Configure per-model fallback chains so Empryo swaps to a backup model on 429s, 529 overloads, and timeouts. Learn what counts as transient and how to tune
- [Plan before you code](https://empryo.com/docs/recipes/plan-mode.md): Use Empryo plan mode for multi-file refactors: the agent researches read-only, writes a light or full plan, waits for approval, then executes step by step.
- [Rename & refactor](https://empryo.com/docs/recipes/rename-refactor.md): Rename symbols across every file, move symbols and files without breaking imports, extract functions, and organize imports using Empryo's LSP-backed refactors.
- [Use multiple models](https://empryo.com/docs/recipes/task-router.md): Assign cheap models to exploration and strong models to code edits with Empryo's task router slots, cutting token cost without losing quality on real work.
- [Test, lint, commit](https://empryo.com/docs/recipes/test-and-commit.md): Run tests, lint, and typecheck through the agent with auto-detected toolchains, then commit with a clean conventional-commit message after pre-commit checks
- [CLI reference](https://empryo.com/docs/reference/cli.md): Every empryo command-line flag, subcommand, exit code, and environment variable — the interface that launches the TUI, desktop app, and headless runs.
- [Command reference](https://empryo.com/docs/reference/commands.md): Full reference for Empryo's 125+ slash commands and keybindings, grouped by session, git, models, intelligence, editor, tabs, and system.
- [Configuration](https://empryo.com/docs/reference/configuration.md): Every Empryo config field: default model, thinking, Genome graph, task router, agent features, compaction, key storage, instruction files, and env vars.
- [Presets](https://empryo.com/docs/reference/presets.md): Load shareable Empryo config bundles by name, path, or URL to sync task router, agent features, theme, and providers across machines and teams.
- [Sponsors & Backers](https://empryo.com/docs/sponsors.md): Sponsors and backers funding Empryo, a graph-powered AI coding agent, with GitHub Sponsors tiers from $5 to $1,000 per month and one-time options.
- [Desktop app](https://empryo.com/docs/surfaces/desktop.md): The Empryo desktop app: a native multi-tab client with the living Genome map, Monaco preview, visual diffs, embedded terminals, and drawer-based controls.
- [Headless & output formats](https://empryo.com/docs/surfaces/headless.md): The contract for empryo --headless: plain, JSON, and JSONL event output shapes, multi-turn chat over stdin, exit codes, and what the runner disables.
- [Surfaces](https://empryo.com/docs/surfaces/overview.md): One engine, three surfaces: the interactive terminal UI, the native desktop app, and the headless CLI. The same Genome, tools and sessions under each.
- [Terminal UI](https://empryo.com/docs/surfaces/tui.md): The interactive Empryo TUI: keybindings, the chat input, tabs, panels, the status bar, forge modes, mouse support, and the embedded Neovim editor.
- [Voice input](https://empryo.com/docs/surfaces/voice-input.md): Speak prompts instead of typing across Empryo's TUI, desktop app, and headless CLI. Record locally, transcribe with your provider key, then review or send.
- [Addons](https://empryo.com/docs/tools/addons.md): Install Empryo's optional addons, CLIProxyAPI for subscription-based LLM routing and a bundled Neovim, on demand from the CLI with PATH-aware management.
- [Agents](https://empryo.com/docs/tools/agents.md): Define specialized subagents as markdown files. Empryo reads them from project and global scopes, including your existing .claude and .opencode dirs.
- [AST editing](https://empryo.com/docs/tools/ast-edit.md): AST-native editing for TypeScript and JavaScript. Address symbols by kind and name, mutate them structurally with 65+ atomic operations across three tiers.
- [The browser](https://empryo.com/docs/tools/browser.md): One browser tool on every surface: an embedded Chromium with audits on desktop, a headless webview in the TUI and in scripts. Local addresses by default.
- [Checkpoints](https://empryo.com/docs/tools/checkpoints.md): Git-backed undo, redo, and branching for agent actions. Browse past checkpoints, rewind files and conversation, or branch from any point with per-tab history.
- [Editor & Neovim](https://empryo.com/docs/tools/editor.md): Embedded Neovim with your config, plugins, and LSP, plus 19 editor tool actions for reading, editing, navigation, and LSP operations from the AI agent.
- [Hearth - Remote control [experimental]](https://empryo.com/docs/tools/hearth.md): Control your running Empryo from Telegram or Discord. Send prompts, approve edits with inline buttons, keep code on your host. Experimental, expect rough edges.
- [Hooks](https://empryo.com/docs/tools/hooks.md): Run shell commands on agent lifecycle events: before/after tool calls, session start/stop, compaction. Wire-compatible with Claude Code settings.json hooks.
- [Languages](https://empryo.com/docs/tools/languages.md): Switch the interface language live with /language, full RTL support on desktop, live translation updates between releases, and a public repo translators can PR.
- [LSP servers](https://empryo.com/docs/tools/lsp.md): Install, manage, and inspect LSP servers from the TUI. Mason-backed installer with 576+ packages, dual Neovim and standalone backends, no config edits.
- [MCP Servers](https://empryo.com/docs/tools/mcp.md): Connect Empryo to Model Context Protocol servers over stdio, Streamable HTTP, or SSE. Namespaced tools, auto-restart, retry with backoff, and headless support.
- [Memory](https://empryo.com/docs/tools/memory.md): Empryo's cross-session memory stores preferences, decisions, and gotchas, then injects up to 3 relevant ones per turn. Per-project or global, with soft-delete
- [Cells & Morphs (experimental)](https://empryo.com/docs/tools/morphs.md): Morph the desktop app with one JSON file: panels you work in (notes, boards, timers, feeds), chips, commands, chords and rules for the agent. No code to write.
- [Project tool](https://empryo.com/docs/tools/project.md): Empryo auto-detects your toolchain across 21 ecosystems and runs the right lint, typecheck, test, and build commands, with code-enforced pre-commit checks and
- [Python kernel](https://empryo.com/docs/tools/python-kernel.md): RLM mode: a persistent Python interpreter on the tool belt. State survives between calls, IPython when available, opt-in and off by default.
- [Routines](https://empryo.com/docs/tools/routines.md): Repo-scoped cron routines: /schedule runs headless agent turns on a cron pattern while Empryo is open. Nightly reviews and recurring checks, no system service.
- [Sessions](https://empryo.com/docs/tools/sessions.md): Empryo auto-saves every conversation as crash-resilient JSONL. Resume by prefix, export to markdown or JSON, branch with tabs, and restore checkpoint git tags.
- [Skills](https://empryo.com/docs/tools/skills.md): Extend Empryo's agent with markdown SKILL.md files. Browse skills.sh, install per project or global scope, and toggle skills into the system prompt per session.
- [Floating Terminals](https://empryo.com/docs/tools/terminals.md): Run up to 5 floating terminal panels in Empryo, powered by Ghostty's GPU-accelerated emulator. Watch logs and REPLs alongside chat, persisting across context
- [Themes](https://empryo.com/docs/tools/themes.md): Switch between 44 builtin terminal themes with live preview, or build custom JSON themes that hot-reload as you edit, plus transparency and opacity controls.
- [Troubleshooting](https://empryo.com/docs/troubleshooting.md): Fix common Empryo issues: missing models, rate limits, Nerd Font icons, LSP and Neovim setup, full context, Copilot tokens, and Hearth daemon errors.