Server: RAM 74% • CPU 24%
Search Chats, Projects & Artifacts
Projects & Autonomous Workspaces
Store coding, design, and reasoning tasks with persistent memory and local files.
Scheduled Tasks & Automation
Iris executes autonomous offline research, code audits, and summaries on your device.
Document
Artifact View
Developer Manifesto
Why We Built Iris
"Artificial intelligence should not be a privileged utility metered behind a monthly toll booth. It belongs in the hands of every coder, dreamer, student, and creator in their own home."
— A Developer's MissionThe Vision
As a software developer and builder, I watched the AI ecosystem rapidly centralize into corporate monopolies charging $20–$200/month subscriptions, gating access, tracking user data, and shutting out independent creators.
I built Iris with a singular conviction: to make sure that every person has access to state-of-the-art intelligence on their own terms. Whether you have an offline laptop in a classroom or a cloud workstation, Iris gives you the power to invent, build, and reason without a paywall.
01
100% Free & Offline
Run open-weights models (DeepSeek R1, Llama 3.2, Qwen 2.5) directly on your device CPU/GPU. No subscriptions, no internet required, zero cloud latency.
02
Bring Your Own Keys (BYOK)
When you need massive frontier power (Claude 3.7, GPT-4o, Gemini 2.5), connect your own direct API keys at true wholesale API cost. Zero middleman markup.
03
Sovereign & Private
Your conversations, local documents, and code stay on your machine. Iris has no telemetry, no surveillance, and never uses your thinking to train models.
Ready to create something extraordinary?
Explore 26+ offline and frontier models now.
Frontier Cloud & BYOK Key Manager
Connect your own direct API keys to unlock 26+ cloud models (Claude 3.7, GPT-4o, Gemini 2.5, DeepSeek R1, Groq). Zero markup.
🔒
Keys are stored strictly in your browser (and your private encrypted Iris account). They are never shared or logged.
OpenAI
GPT-4o, o1, o3-mini
Anthropic
Claude 3.7 Sonnet, 3.5 Haiku
Google Gemini
Gemini 2.5 Flash, 2.5 Pro, 2.0
DeepSeek Cloud
DeepSeek V3, DeepSeek R1
Groq Cloud (Ultra Fast)
Llama 3.3 70B, 3.1 8B, Mixtral
Mistral AI
Mistral Large 2, Codestral
OpenRouter (Universal Router)
Unified access to 100+ AI models
APIs — Custom OpenAI-Compatible Endpoints
Bring your own custom AI endpoints (vLLM, LM Studio, Ollama, Together AI, xAI, Perplexity, LiteLLM, corporate gateways).
Quick Provider Presets:
+
Add Custom OpenAI-Compatible Endpoint
Configured Custom Endpoints
0 configuredWelcome to Iris
Create a free account to secure your workspace, sync BYOK keys, and prevent unauthorized abuse.
Local Models Hub
100% offline GGUF quantized models running directly on your CPU/GPU with zero cloud latency.
Iris Live Voice
Listening... Speak naturally to Iris.
● 100% Offline Local Whisper
Listening to microphone... Speak to Iris now.