dulus 3.6.1

Use IA as an agent without an API key. $0. Harvest your browser session for Gemini (guest, no login), Claude.ai, Claude Code, Kimi.com, Qwen, DeepSeek — drive them like an API, no per-token billing. Plus LiteLLM (100+ paid providers), local models via Ollama, /lang in 34 languages, Mesa Redonda, voice, OCR, MemPalace, embedded sandbox OS.

▲ DULUS

pip install dulus → press Enter → working IA in 30 seconds. Try it.

SET /sticky_input ON since the first run for the best experience!

The Dulus (Cigua Palmera) — Dominican national bird. Named after the bird, not the rocket.

Quick Start · Models · Features · Permissions · MCP · Plugins

🌐 Multilingual · English · Español · Français · 中文 · 日本語 · 한국어 · Português · Русский · العربية

Official X / creator handle: @KevRojo — that's me, the only contributor to this repo. Any other account claiming to be Dulus is a copycat.

$DULUS AI token Solana): 9R8rrjXxcfQPmLTCLhmVpjr2uesjjkcgkinE6Lwdpump

---

pip install dulus

🌐 Visit the Dulus website →
_{The site covers features, demos, and details not documented in this README.}

🔥 What's new

Full changelog: docs/news.md · Inside the REPL: /news

• 📡 dulus.work — announced. Kevin finally revealed the final piece of the Dulus vision — the one he'd been keeping secret (even from Dulus itself 😂). Every agent ships updates. Nobody ships THIS. Already in motion. Stay tuned. 🦅 → dulus.work

• Dulus Agent — Telegram communities. Dulus evolved from a CLI tool into a live AI agent inside Telegram groups. The first fully autonomous AI moderating and conversing in real communities — not a bot, not a filter, a real agent. Groups pay in $DULUS to activate him. We host Dulus for them — that's the business model. The CLI stays free forever. This is the paid layer. 🦅🇩🇴

• v0.2.93 — IA WITHOUT AN API KEY, FIRST-RUN. 🦅 The welcome wizard now offers, by default, to open Gemini in a browser and capture its guest session — no Google login, no API key, no credit card. From pip install to working IA in 30 seconds. Same flow works for Claude.ai / Kimi.com / Qwen / DeepSeek if you have those accounts. This was Dulus's ace under the sleeve. Now it's the front door.

• v0.2.92 — /lang command. 34 ISO codes + free-form descriptors. /lang zh, /lang ja, /lang pt-br, /lang "speak as my gym tutor", /lang "tigre de calle dominicano", /lang "Yoda". The model role-plays the voice across the whole session. Triggered after Doubao (China's biggest IA assistant) started referring traffic to the repo.

• v0.2.91 — CORS on the daemon. The Android Sandbox APK now connects every app live — same daemon, finally cross-origin reachable.

• v0.2.89 — Sandbox embedded INSIDE the desktop GUI. Click 🦅 Sandbox in the GUI Web tab → the Dulus OS renders inside the content frame via pywebview subprocess + Win32 SetParent reparent. Not a popup. Inside the frame.

• v0.2.89 — Local OCR shipped first-class. /ocr + ExtractTextFromImage tool, WebBridgeScreenshot auto-OCRs, /img sends image + verbatim OCR text together. Zero vision-model tokens for receipts / code screenshots / error stacks.

• v0.2.89 — kepano/obsidian-skills bundled. Dulus writes Obsidian-flavored markdown by default. Open ~/.dulus/memory/ as an Obsidian vault → instant graph view.

• v0.2.88 — LiteLLM gateway: one provider entry, 100+ backends. OpenRouter, Groq, Together, Bedrock, Vertex, Cohere, Perplexity, xAI, Mistral, Fireworks, Anyscale, Replicate, Azure, DeepInfra — the welcome wizard auto-installs the package and asks for the right per-backend key.

• v0.2.85 — Slim wheel 11.4 MB → 2.5 MB. Smaller than the original baseline.

What is this

Dulus Reduce your IA costs by 60% parsing webchats and claude-code directly. Write poetry while Anthropic only see text. Use claude-code as an API without the new 'extra-usage' wall <3

Another reminder of a Dulus magic spell: Wanna get stock prices, history , etc?

/plugin install yfinance@https://github.com/ranaroussi/yfinance

them: /plugin reload

dulus get the prices of NVDA, TSLA, SP500:

Be creative!!!

Dulus adapt any python repository <3

Quick Start

Windows — no Python needed (1-click):

Download Dulus-Free-3.3.3-x64.msi → double-click and fly. Embedded Python, every library, and the desktop GUI bundled. No pip, no terminal, no admin. ~85 MB.

Everyone else — one command:

pip install dulus
dulus

First run opens a welcome wizard that can get you working IA without an API key — it captures a Gemini guest session from your browser (no login, no card). Or set any key you already have. Zero to flight in 30 seconds.

💡 Set /sticky_input ON on first run for the best experience.

Models

Speak every dialect — switch mid-session with /model:

Kind	Providers
Cloud	Anthropic · OpenAI · Gemini · DeepSeek · Kimi · Qwen · Zhipu · MiniMax
Free tier	14 frontier models via NVIDIA NIM — 40 RPM each, $0. Key at build.nvidia.com
Gateway	100+ backends via LiteLLM — OpenRouter, Groq, Together, Bedrock, Vertex, xAI, Mistral...
Web chats	Use the chats you already pay for: Claude.ai, Gemini, Kimi.com, Qwen, DeepSeek — parsed as providers
Local	Ollama · LM Studio · any OpenAI-compatible endpoint. Runs offline, completely.

Features

The full tour lives on the website — highlights:

• 27 built-in tools — Read/Write/Edit, Bash, Glob/Grep, WebFetch/WebSearch, NotebookEdit, GetDiagnostics, Memory, Tasks, Agents, Skills...
• Sub-agents — typed agents (coder, reviewer, researcher, tester), each in its own git worktree, talking via message passing
• Persistent memory + checkpoints — never lose context. Ever. Open ~/.dulus/memory/ as an Obsidian vault
• Dulus OS — not a CLI, a workstation for the agent (sandbox desktop in your browser)
• Voice — offline Whisper STT, no API key, no cloud. Wake words included
• Brainstorm Mode — multi-persona AI debate. Council of ghosts
• SSJ Developer Mode — ten chained workflow shortcuts behind one keystroke: refactor → review → test → commit → ship
• Telegram bridge — slash commands, vision, and voice from your phone
• Local OCR — /ocr, /img with verbatim text extraction, zero vision tokens
• Plan mode, webchat UI, desktop GUI, task manager... — see the site

Permissions

Three modes, switchable any time with /permissions [mode]:

Mode	Behavior
auto (default)	Safe tools run free; risky actions (writes, shell) ask first
manual	Everything asks
accept-all	Never asks — YOLO. Also via dulus --accept-all

Answering a at any prompt upgrades the session to accept-all.

MCP

Drop a .mcp.json in your project — every server registers instantly as mcp__server__tool. stdio, SSE, and HTTP transports.

/mcp                        # list servers and their tools
/mcp add <name> <cmd>       # add a stdio server
/mcp reload                 # reconnect all

Plugins

The Auto-Adapter onboards any Python repo as native tools — zero manifest, hot-reload, no restart:

/plugin install yfinance@https://github.com/ranaroussi/yfinance
/plugin reload

Then just ask: "dulus get the prices of NVDA, TSLA, SP500". Be creative — 800+ ready-made skills via Composio too.

🔒 Privacy & anonymous telemetry (opt-in)

On first launch Dulus asks once if you'd like to share anonymous usage statistics. Nothing is sent unless you say yes.

Collected (if you opt in)	NEVER collected
Event names (session_start, tool_used, model_selected)	Prompts or responses
Dulus version, OS name, Python version	File contents or paths
Provider/model names (e.g. gemini)	API keys or tokens
A random anonymous ID generated on your machine	Emails, usernames, IPs (geo disabled)

Data goes to Mixpanel (event analytics). The full implementation is ~140 lines of readable code in analytics.py — audit it yourself.

Opt out anytime:

/config telemetry=off        # inside Dulus
DULUS_TELEMETRY=0            # environment variable