English · 中文
Mente is a unified AI agent for coding, automation, gateway workflows, and long-term memory. It creates skills from experience, improves them during use, nudges itself to persist knowledge, searches its own past conversations, and builds a deepening model of who you are across sessions. Run it on a $5 VPS, a GPU cluster, or serverless infrastructure that costs nearly nothing when idle. It's not tied to your laptop — talk to it from Telegram while it works on a cloud VM.
This branch also completes a product-surface consolidation pass:
- External surface is uniformly Mente across CLI, gateway progress, messaging, and user-facing replies.
- Internal execution still runs on the Codex-backed executor — the rename is presentation-layer cleanup, not a runtime downgrade.
- Gateway progress is visible again with Mente-facing step names while preserving detailed command/tool activity.
- Config/admin operations are now explicit through a dedicated Mente skill for API keys, provider auth,
.env,config.yaml, and gateway restart rules.
Use any model you want — Nous Portal, OpenRouter (200+ models), NVIDIA NIM (Nemotron), Xiaomi MiMo, z.ai/GLM, Kimi/Moonshot, MiniMax, Hugging Face, OpenAI, or your own endpoint. Switch with mente model — no code changes, no lock-in.
| A real terminal interface | Full TUI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output. |
| Lives where you do | Telegram, Discord, Slack, WhatsApp, Signal, and CLI — all from a single gateway process. Voice memo transcription, cross-platform conversation continuity. |
| A closed learning loop | Agent-curated memory with periodic nudges. Autonomous skill creation after complex tasks. Skills self-improve during use. FTS5 session search with LLM summarization for cross-session recall. Honcho dialectic user modeling. Compatible with the agentskills.io open standard. |
| Scheduled automations | Built-in cron scheduler with delivery to any platform. Daily reports, nightly backups, weekly audits — all in natural language, running unattended. |
| Delegates and parallelizes | Spawn isolated subagents for parallel workstreams. Write Python scripts that call tools via RPC, collapsing multi-step pipelines into zero-context-cost turns. |
| Runs anywhere, not just your laptop | Six terminal backends — local, Docker, SSH, Daytona, Singularity, and Modal. Daytona and Modal offer serverless persistence — your agent's environment hibernates when idle and wakes on demand, costing nearly nothing between sessions. Run it on a $5 VPS or a GPU cluster. |
| Research-ready | Batch trajectory generation, Atropos RL environments, trajectory compression for training the next generation of tool-calling models. |
curl -fsSL https://raw.githubusercontent.com/chemany/Mente/main/scripts/install.sh | bashWorks on Linux, macOS, WSL2, and Android via Termux. The one-click installer is release-pinned by default and can also bootstrap a matching vendored Codex runtime from local/offline assets via --runtime-artifact-manifest and --runtime-wheel.
npm install -g mente-agent
menteThe npm package is intentionally thin. It publishes only the launcher and installer scripts, then bootstraps the full Mente runtime on first run. By default the bootstrapper installs from the repo's main branch, and you can force a tagged release with MENTE_BOOTSTRAP_RELEASE=<tag> mente. It does not publish your local .env, auth.json, ~/.mente, ~/.hermes, sessions, logs, or other machine-local state.
At the moment, the repository is ready for npm publish but not yet live on the npm registry. Until the first public npm release is published, use Option 1 above. Once the package is published, the npm install -g mente-agent flow becomes the primary one-line install path.
Release operators can use the short npm runbook here: docs/releasing/npm.md.
Android / Termux: The tested manual path is documented in the Termux guide. On Termux, Mente installs a curated
.[termux]extra because the full.[all]extra currently pulls Android-incompatible voice dependencies.Windows: Native Windows is not supported. Please install WSL2 and run the command above.
Developers / source checkouts: use
./setup-hermes.shafter cloning the repo manually. That path is for editable development, not the frozen end-user release install.
After installation:
source ~/.bashrc # reload shell (or: source ~/.zshrc)
mente # start chatting!mente # Interactive CLI — start a conversation
mente model # Choose your LLM provider and model
mente tools # Configure which tools are enabled
mente config set # Set individual config values
mente gateway # Start the messaging gateway (Telegram, Discord, etc.)
mente setup # Run the full setup wizard (configures everything at once)
mente claw migrate # Migrate from OpenClaw (if coming from OpenClaw)
mente update # Update to the latest version
mente doctor # Diagnose any issuesThis README tracks the current Mente packaging and runtime direction:
- One current install command for GitHub visitors:
curl -fsSL https://raw.githubusercontent.com/chemany/Mente/main/scripts/install.sh | bash - One intended npm install command after publish:
npm install -g mente-agent - One visible agent identity: user-facing replies and progress now present as
Mente - Same execution depth under the hood: complex coding and tool work still route through the Codex-backed executor
- Safer operations surface: packaging is whitelist-based, and config/admin actions now have explicit handling for API keys, provider auth, and restart boundaries
If you are evaluating Mente from GitHub, the practical model is:
- Install Mente with the direct installer.
- Launch
mente. - Let the bootstrap flow finish preparing the full runtime.
- Use Mente normally from CLI or gateway surfaces.
Mente has two entry points: start the terminal UI with mente, or run the gateway and talk to it from Telegram, Discord, Slack, WhatsApp, Signal, or Email. Once you're in a conversation, many slash commands are shared across both interfaces.
| Action | CLI | Messaging platforms |
|---|---|---|
| Start chatting | mente |
Run mente gateway setup + mente gateway start, then send the bot a message |
| Start fresh conversation | /new or /reset |
/new or /reset |
| Change model | /model [provider:model] |
/model [provider:model] |
| Set a personality | /personality [name] |
/personality [name] |
| Retry or undo the last turn | /retry, /undo |
/retry, /undo |
| Compress context / check usage | /compress, /usage, /insights [--days N] |
/compress, /usage, /insights [days] |
| Browse skills | /skills or /<skill-name> |
/<skill-name> |
| Interrupt current work | Ctrl+C or send a new message |
/stop or send a new message |
| Platform-specific status | /platforms |
/status, /sethome |
For the full command lists, see the CLI guide and the Messaging Gateway guide.
All documentation lives at chemany.github.io/Mente/docs:
| Section | What's Covered |
|---|---|
| Quickstart | Install → setup → first conversation in 2 minutes |
| CLI Usage | Commands, keybindings, personalities, sessions |
| Configuration | Config file, providers, models, all options |
| Messaging Gateway | Telegram, Discord, Slack, WhatsApp, Signal, Home Assistant |
| Security | Command approval, DM pairing, container isolation |
| Tools & Toolsets | 40+ tools, toolset system, terminal backends |
| Skills System | Procedural memory, Skills Hub, creating skills |
| Memory | Persistent memory, user profiles, best practices |
| MCP Integration | Connect any MCP server for extended capabilities |
| Cron Scheduling | Scheduled tasks with platform delivery |
| Context Files | Project context that shapes every conversation |
| Architecture | Project structure, agent loop, key classes |
| Contributing | Development setup, PR process, code style |
| CLI Reference | All commands and flags |
| Environment Variables | Complete env var reference |
If you're coming from OpenClaw, Mente can automatically import your settings, memories, skills, and API keys.
During first-time setup: The setup wizard (mente setup) automatically detects ~/.openclaw and offers to migrate before configuration begins.
Anytime after install:
mente claw migrate # Interactive migration (full preset)
mente claw migrate --dry-run # Preview what would be migrated
mente claw migrate --preset user-data # Migrate without secrets
mente claw migrate --overwrite # Overwrite existing conflictsWhat gets imported:
- SOUL.md — persona file
- Memories — MEMORY.md and USER.md entries
- Skills — user-created skills →
~/.hermes/skills/openclaw-imports/ - Command allowlist — approval patterns
- Messaging settings — platform configs, allowed users, working directory
- API keys — allowlisted secrets (Telegram, OpenRouter, OpenAI, Anthropic, ElevenLabs)
- TTS assets — workspace audio files
- Workspace instructions — AGENTS.md (with
--workspace-target)
See mente claw migrate --help for all options, or use the openclaw-migration skill for an interactive agent-guided migration with dry-run previews.
We welcome contributions! See the Contributing Guide for development setup, code style, and PR process.
Quick start for contributors — clone and go with setup-hermes.sh:
git clone https://github.com/chemany/Mente.git
cd Mente
./setup-hermes.sh # installs uv, creates venv, installs .[all], symlinks ~/.local/bin/mente
./mente # auto-detects the venv, no need to `source` firstManual path (equivalent to the above):
curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv venv --python 3.11
source venv/bin/activate
uv pip install -e ".[all,dev]"
scripts/run_tests.shRL Training (optional): The RL/Atropos integration (
environments/) ships via theatroposlibandtinkerdependencies pulled in by.[all,dev]— no submodule setup required.
- 💬 Discord
- 📚 Skills Hub
- 🐛 Issues
- 🔌 HermesClaw — Community WeChat bridge: Run Mente and OpenClaw on the same WeChat account.
MIT — see LICENSE.
Built for the Mente project.