Skip to main content

Subnaut

The agent for Bittensor, from Subnaut Labs. It reads subnets, metagraphs and stake straight from the chain, runs btcli with your approval on every signature, and ships skills for miners, validators and subnet owners. A built-in learning loop turns what it works out into skills it keeps and improves, and it remembers across sessions.

Get started →Bittensor chain toolsExplore bot modeDesktop appBrowse skills

Install​

Windows or macOS​

Install the command line with the installer for your platform below, then run subnaut desktop: it builds the desktop app and launches it against the same install. There is no separate desktop download; the release channel publishes the installers and source archives.

Without Subnaut Desktop:​

For a command-line only install without Subnaut Desktop, run:

Linux / macOS / WSL2 / Android (Termux)​

curl -fsSL https://tryimagent.com/install.sh | bash

Windows (native)​

Run in powershell:

iex (irm https://tryimagent.com/install.ps1)

See the full Installation Guide for what the installer does, the per-user vs root layout, and Windows-specific notes. For the complete platform support matrix, see Platform Support.

What is Subnaut Agent?​

It's not a coding copilot tethered to an IDE or a chatbot wrapper around a single API. It's an autonomous agent that gets more capable the longer it runs. It lives wherever you put it — a $5 VPS, a GPU cluster, or serverless infrastructure (Daytona, Modal) that costs nearly nothing when idle. Talk to it from Telegram while it works on a cloud VM you never SSH into yourself. It's not tied to your laptop.

🚀 InstallationInstall in 60 seconds on Linux, macOS, WSL2, native Windows, Nix & NixOS or Android
📖 Quickstart TutorialYour first conversation and key features to try
🗺️ Learning PathFind the right docs for your experience level
τ Bittensor Chain ToolsSubnets, metagraphs and accounts, read straight from the chain
⚙️ ConfigurationConfig file, providers, models, and options
💬 Messaging GatewaySet up Telegram, Discord, Slack, and webhooks
🤖 Bot ModeNamed Bots with their own model, memory, skills, routines, and chats
🔧 Tools & Toolsets70 built-in tools and how to configure them
🧠 Memory SystemPersistent memory that grows across sessions
📚 Skills SystemProcedural memory the agent creates and reuses
🔌 MCP IntegrationConnect to MCP servers, filter their tools, and extend Subnaut safely
🧭 Use MCP with SubnautPractical MCP setup patterns, examples, and tutorials
🎙️ Voice ModeReal-time voice interaction in CLI, Telegram, Discord, and Discord VC
🗣️ Use Voice Mode with SubnautHands-on setup and usage patterns for Subnaut voice workflows
🎭 Personality & SOUL.mdDefine Subnaut's default voice with a global SOUL.md
📄 Context FilesProject context files that shape every conversation
🔒 SecurityCommand approval, authorization, container isolation
💡 Tips & Best PracticesQuick wins to get the most out of Subnaut
🏗️ ArchitectureHow it works under the hood
❓ FAQ & TroubleshootingCommon questions and solutions

Key Features​

  • Built for Bittensor — Read-only chain tools for subnets, metagraphs, balances and stake; btcli that asks before every signature and never touches key material; bundled skills for miners, validators and subnet owners
  • A closed learning loop — Agent-curated memory with periodic nudges, autonomous skill creation, skill self-improvement during use, FTS5 cross-session recall with LLM summarization, and Honcho dialectic user modeling
  • Runs anywhere, not just your laptop — 7 terminal backends: local, Docker, SSH, Daytona, Singularity, Modal, Vercel Sandbox. Daytona and Modal offer serverless persistence — your environment hibernates when idle, costing nearly nothing
  • Lives where you do — CLI and TUI, the desktop app, your editor over ACP, and Telegram, Discord, Slack, webhooks and an OpenAI-compatible API server, all from one gateway
  • Bring your own model — 36 providers in the subnaut model picker, including Chutes (decentralised inference on Bittensor subnet 64), OpenRouter, OpenAI and Anthropic, or any OpenAI-compatible endpoint
  • Scheduled automations — Built-in cron with delivery to any platform
  • Bot Mode — Build a durable team of specialist Bots that work together in group chats and through @mentions
  • Delegates & parallelizes — Spawn isolated subagents for parallel workstreams. Programmatic Tool Calling via execute_code collapses multi-step pipelines into single inference calls
  • Open standard skills — Compatible with agentskills.io. Skills are portable, shareable, and community-contributed via the Skills Hub
  • Full web control — Search, extract, browse, vision, image generation, TTS — each backed by the provider and key you choose
  • MCP support — DeepWiki and Context7 are connected by default; add catalog servers with subnaut mcp install <name> or connect any other MCP server
  • Research-ready — Trajectory export in ShareGPT format

For LLMs and coding agents​

Machine-readable entry points to this documentation:

  • /llms.txt — curated index of every doc page with short descriptions. ~29 KB, safe to load into an LLM context.
  • /llms-full.txt — every doc page concatenated into a single markdown file for one-shot ingestion. ~3.5 MB.

Both are generated fresh on every deploy.