One install. Every model. Every channel. Full autonomy.
The harness that treats the LLM as just the inference call — memory, tools, sandbox, orchestration and learning all live above it.
curl -fsSL https://raw.githubusercontent.com/gabrielvuksani/wotann/main/install.sh | bash
Runs entirely locally by default via Ollama + free tiers. Drop an API key to upgrade selectively — no lock-in, no telemetry unless you opt in.
Most agent frameworks lock you to a vendor, degrade quietly when a model can't tool-call, and hand you a single UI. WOTANN inverts that. It lives at the layer above inference so your choice of model is always a drop-in.
Anthropic · OpenAI · Codex OAuth · Copilot · Gemini · Groq · Ollama · Bedrock · Vertex · HuggingFace · Azure · DeepSeek · Perplexity · xAI · Together · Fireworks · SambaNova · Cerebras · Mistral — one router, automatic fallback chaining.
Tool-calling, vision (OCR + accessibility tree), and extended thinking are synthesized for any model that lacks them natively. A 3B local Gemma gets the same scaffolding as Claude Opus.
SQLite + FTS5 + vector + graph-RAG + episodic + semantic + temporal + working + archival. One search API across all layers; context replay under an explicit token budget.
Verify-fix loop, self-healing provider fallback, 8-strategy escalation, doom-loop detector. Every completion ships a proof bundle — tests, typecheck, diff, optional screenshots.
CLI · TUI · Tauri 2 desktop · iOS · Apple Watch · CarPlay · Widgets · Live Activities · Share Extension · Siri Intents. Same session, same memory, same agent.
Ollama + Groq + Gemini 2.5 Flash free tier give the full experience. No paywall to the good features, no quiet model downgrades when you're on a subscription.
Telegram · Slack · Discord · Signal · WhatsApp · iMessage · Teams · Matrix · Email · SMS · Webhooks · GitHub bot · IDE bridge · IRC · Google Chat · and more.
Guardian post-response LLM-as-judge. Self-consistency voting. Council deliberation with chairman synthesis. Dream pipeline nightly consolidation at 02:00–04:00.
Telemetry is opt-out and triple-gated (DO_NOT_TRACK, WOTANN_NO_TELEMETRY, sentinel file). Runs locally by default. No phone-home. Session token auth on every RPC call.
| Most agents | WOTANN | |
|---|---|---|
| Provider | Locked to one vendor | 19 providers via one router · Anthropic & OpenAI subscriptions or API keys |
| Capabilities | Tool-calling · vision · thinking are gated by model | All three injected for any model — even a 3B local Gemma |
| Free tier | An afterthought | First-class — Ollama + Groq + Gemini 2.5 Flash give the full experience |
| Memory | Short-lived prompt context | 8-layer persistent — SQLite + FTS5 + vector + graph-RAG + episodic + temporal + provenance |
| Surfaces | One UI | 7 surfaces — CLI · TUI · Desktop · iOS · Watch · CarPlay · 25+ channels |
| Autonomy | Vibes | Proof bundles (tests + typecheck + diff + screenshots) on every completion |
| Lock-in | Total | Zero — swap models mid-session, export all state |
Same conversation, same agent, same history — wherever you touch it.
78+ commands. wotann init, wotann engine, wotann autonomous, wotann arena, wotann link, wotann cost.
Ink-rendered REPL · Obsidian Precision theme · voice push-to-talk · slash commands · OSC 133 blocks.
Chat · Editor (Monaco) · Workshop · Exploit · Trust · Integrations · Command palette ⌘K · Computer Use panel.
SwiftUI main app · work tab with filter pills · onboarding wizard · pairing via PIN or Bonjour.
Quick actions (tests / build / lint / approve / voice). Hands-free dispatch with list + detail templates.
Cost + agent status widgets · Dynamic Island · Share Extension · AskWOTANN · CheckCost · EnhancePrompt.
Telegram · Slack · Discord · Signal · WhatsApp · iMessage · Teams · Matrix · Email · SMS · Webhooks · GitHub bot · IDE bridge · IRC · Google Chat · Mastodon · WeChat · LINE · …
15s tick · cron · heartbeat · event triggers · session-token RPC. Runs at ~/.wotann/kairos.sock.
curl -fsSL https://raw.githubusercontent.com/gabrielvuksani/wotann/main/install.sh | bashDetects your platform, downloads the SEA binary (Node is bundled — no system Node needed), places it on your $PATH.
npm install -g wotannRequires Node ≥ 20. Ships as an ESM package with a single wotann binary. Works everywhere npm works.
brew tap gabrielvuksani/wotann
brew install wotannFormula is in Formula/wotann.rb with the v0.4.0 macOS arm64 SHA. Tap repo publication lands in the next release.
wotann init # initialize a workspace
wotann engine # start the always-on daemon
wotann # launch the TUIThat's the whole setup. Drop API keys later to upgrade selectively:
export ANTHROPIC_API_KEY=sk-ant-...
export OPENAI_API_KEY=sk-...
export GEMINI_API_KEY=AIza...
# Or stay free
ollama pull qwen2.5-coder:7b
wotann # auto-detects OllamaWOTANN itself is free and open source under the MIT license. You pay only the model providers you choose to use — and all of them are optional.
$0
Pay providers directly
No phone-home. No analytics pings. No crash reporting without consent. Telemetry is opt-out and triple-gated.
Single command on a fresh Mac or Linux box — five minutes to your first autonomous task.