OmniCode

The AI-Native IDE with a Built-In Provider Control Center.
Connect 169+ AI providers. Route across free & paid models. Zero config.

🚀 Quick Start • 🔌 Providers • 🎯 Combos • 📊 Cost Savings • 🛡️ OmniProxy • 📸 Screenshots • 📖 Docs

💥 The Promise • 🤔 Why • 🏆 What Sets Apart • 🔒 Security • 🏗️ Architecture • 🤝 Contributing

💥 The Promise

One IDE. 169+ providers. Every AI model you need — and a free coding agent that never stops.

OmniCode is an intelligent development environment that centralizes and streamlines every AI interaction directly inside your editor. Stop juggling browser dashboards, dead API keys, and surprise bills. OmniCode brings it all home.


🚫 Never hit limits	💸 $0 to start	📊 Full cost visibility
Auto-fallback across providers. Quota out? Next model takes over — zero downtime.	Free-tier OAuth providers (Kiro, Qoder, GitHub Copilot, Gemini CLI). No card needed.	Real-time dashboards for token usage, per-provider costs, and model-level spend.
🔌 Every model works	🧩 One unified picker	🛡️ Production-grade
169+ providers — Claude, GPT, Gemini, DeepSeek, Grok, Qwen — through one interface.	All models synced into the native editor picker. No separate UI or fragmented experience.	Semantic caching, combo routing, budget controls, PII guards, encryption at rest.

🤔 Why OmniCode?

Stop juggling 10 dashboards, dead API keys, and surprise bills.

❌ The daily pain	✅ How OmniCode fixes it
📉 Subscription quota expires unused every month	Maximize subscriptions — track quota, use every token before reset
🛑 Rate limits stop you mid-coding	Auto-fallback combos — route to the next provider in milliseconds
💸 Expensive APIs ($20–50/mo per provider)	Free coding agent — route routine tasks to free-tier models at $0
🧰 Each AI tool wants its own setup	One IDE, one picker — every model in one unified dropdown
🔐 API keys scattered across .env files	Encrypted secret storage — AES-256-GCM at rest, never in source control
📊 No visibility into what you're spending	Cost & analytics dashboards — per-request, per-model, per-provider breakdowns

┌──────────────────────────────────────────────────────────┐
│       OmniCode IDE  (Native Chat & AI Agent Interface)     │
└─────────────────────────┬──────────────────────────────────┘
                          │ Embedded OmniProxy Runtime
                          ▼
┌──────────────────────────────────────────────────────────┐
│              OmniProxy — Smart Router & Cache               │
│  Semantic caching · Combo routing · Budget controls         │
│  PII guards · Encrypted secrets · Cost analytics            │
└─────────────────────────┬──────────────────────────────────┘
        ┌─────────────┬────┴────────┬─────────────┐
        ▼ Tier 1      ▼ Tier 2      ▼ Tier 3       ▼ Tier 4
   SUBSCRIPTION     API KEY        FREE TIER      CUSTOM
   Claude Code,     OpenAI,        Kiro, Qoder,   Local LLMs,
   Codex, Copilot   DeepSeek       Gemini CLI     Ollama
   quota out? ───▶  budget hit? ─▶ always on ──▶  your server

🎯 Combos — Multi-Model Routing

A Combo is a chain of models OmniCode routes across automatically. Quota runs out, a provider fails, or costs spike — the combo silently slides to the next model. This is what makes OmniCode unbreakable. 🛡️

⚡ How it works

Configure fallback chains in the OmniProxy dashboard. OmniCode evaluates your rules in real-time:

Combo: "always-on-coding"                    Strategy: priority
  1. cc/claude-opus-4-7      ← OAuth subscription (use it fully)
  2. cx/gpt-5.5              ← Codex OAuth (second subscription)
  3. kr/claude-sonnet-4.6    ← Kiro FREE (unlimited Claude)
  4. qoder/deepseek-v3.2     ← Qoder FREE (always available)
Result: 4 layers of fallback = zero downtime

🔀 Routing strategies

Goal	Strategy
🥇 Drain my subscription before paying	Priority / Fill-first
⚖️ Spread load across accounts	Round-robin
💸 Always cheapest viable model	Cost-optimized
🧠 Best model for complex reasoning	Premium escalation
🆓 Free-first for routine tasks	Free agent routing

🧱 Resilience is built in

Layer	Scope	What it does
🔌 Circuit breaker	Whole provider	Stops hammering a failing provider; auto-probes to recover
💤 Connection cooldown	One account/key	Skips a rate-limited key while others keep serving
💾 Semantic cache	Identical prompts	Returns cached response instantly — cost: $0, latency: ~0ms

🏆 What Sets OmniCode Apart

Feature	OmniCode	Traditional IDEs	Other AI Tools
🔌 Built-in AI providers	169+ (OAuth + API key)	1–2	External only
🆓 Free-tier routing	✅ Automatic	❌	❌
🔀 Multi-model combos	✅ Configurable chains	❌	Rare
📊 Cost & usage dashboards	✅ Native	❌	Separate tool
💾 Semantic caching	✅ Built-in	❌	Rare
🧩 Unified model picker	✅ All models, one dropdown	1 provider	N/A
🔗 Custom endpoints	✅ Any OpenAI-compatible	❌	Limited
🛡️ PII & injection guards	✅ Built-in	❌	Rare
🔐 Encrypted secret storage	✅ AES-256-GCM	Plaintext .env	Varies
🖥️ Full IDE experience	✅ VS Code-based editor	✅	❌ CLI only
📦 Batch testing	✅ Cross-model evaluation	❌	❌
🎨 Image generation	✅ Built-in media tools	❌	Separate

🔌 169+ AI Providers — Unified Model Picker

The most comprehensive provider ecosystem of any AI-native IDE. Connect via OAuth (one-click) or API key — all models appear instantly in the native editor picker.

🔐 OAuth Providers (One-Click Connect)

Provider	Auth Flow	Models	Status
Claude (Anthropic)	Authorization Code + PKCE	Opus 4.7, Sonnet 4.6, Haiku 4.5	✅
Codex (OpenAI)	Authorization Code + PKCE	GPT-5.5, GPT-5.4, GPT-5.3 Codex	✅
Gemini (Google)	Standard OAuth2	Synced via API (all models)	✅
Gemini CLI	Standard OAuth2	Gemini 3.1 Pro, 3 Flash, 3.1 Flash Lite	✅
GitHub Copilot	Device Code Flow	GPT-5.5, Claude Opus 4.7, Gemini 3.1 Pro	✅
Qwen (Alibaba)	Device Code + PKCE	Qwen3 Coder Plus/Flash, Vision	✅
Kimi Coding (Moonshot)	Device Code Flow	Kimi K2.6, K2.6 Thinking	✅
Antigravity (Google Cloud)	Standard OAuth2	Passthrough models	✅
Kiro (AWS)	SSO OIDC / Social Login	Claude Opus 4.7, Sonnet 4.6, Haiku 4.5	✅
Cursor	Token Import	70+ models (GPT, Claude, Gemini, Grok)	✅
Cline	Local Callback Flow	Claude, GPT, Gemini, DeepSeek, Kimi	✅
KiloCode	Custom Device Auth	OpenRouter models, GPT, Claude, Gemini	✅
GitLab Duo	Authorization Code + PKCE	Enterprise models	✅
Amazon Q	AWS Builder ID	AWS models	✅
Qoder AI	API Key / OAuth	Qwen3, Kimi K2, DeepSeek, ROME	✅

🔑 API Key Providers

Provider	Highlights
OpenAI	GPT-5.5, GPT-5.4, GPT-5.4 Mini/Nano, O3
Anthropic	Claude Opus 4.7, Sonnet 4.6, Haiku 4.5
DeepSeek	V4 Pro, V4 Flash
Groq	Llama 3.3 70B, Llama 4 Maverick, Qwen3 32B
xAI	Grok 4.3, Grok 4.20 (Reasoning, Multi Agent)
Mistral	Large 3, Medium 3.5, Small 4, Devstral 2
Together	Llama 3.3 70B 🆓, DeepSeek R1 🆓, Qwen3 235B
Fireworks	Kimi K2.6, MiniMax M2.7, Qwen3.6 Plus
Cerebras	GLM 4.7, GPT OSS 120B
Perplexity	Sonar Deep Research, Reasoning Pro, Pro
Cohere	Command A Reasoning, Vision
NVIDIA NIM	GLM 5.1, Gemma 4, Nemotron 3 Super, DeepSeek V4
SiliconFlow	DeepSeek V3.2, Qwen3 Coder 480B
OpenRouter	Auto (Best Available) — access to 200+ models
HuggingFace	Open-source model catalog
Ollama Cloud	DeepSeek V4, Kimi K2.6, GLM 5.1
MiniMax	M2.7, M2.5, Highspeed variants
GLM / ZAI	GLM 5.1, GLM 5, GLM 5 Turbo

🆓 Free Providers — $0, No Card

Provider	What's Free	Limits
Kiro AI	Claude Opus 4.7, Sonnet 4.6, Haiku 4.5	OAuth, generous limits
Qoder AI	Qwen3, DeepSeek, Kimi K2	Unlimited for supported models
Gemini CLI	Gemini 3.1 Pro, 3 Flash	180K requests/mo free
GitHub Copilot	GPT-5 Mini, Claude Haiku 4.5	Free tier available
Together AI	Llama 3.3 70B, DeepSeek R1 Distill, Llama Vision	Free tier models
NVIDIA NIM	13+ models	~40 RPM free
Cerebras	GLM 4.7, GPT OSS 120B	Free tier with limits
OpenCode	GLM 5.1, Kimi K2.6, DeepSeek V4, MiMo V2.5	Free tier available

🔗 Custom Endpoints

Integrate any OpenAI-compatible endpoint in 3 steps:

Provide a group name
Supply the API key
Enter the Base URL

OmniCode automatically queries the /models endpoint and populates available models into the unified picker.

📊 Cost Savings & Free Coding Agent

OmniCode is built to dramatically reduce API spend through intelligent request handling:

flowchart LR
    A[Incoming Request] --> B{Is it Cached?}
    B -- Yes --> C["Return Cached Response<br/><b>Cost: $0</b>"]
    B -- No --> D{Requires Premium AI?}
    D -- No --> E["Route to Free Agent/Tier<br/><b>Cost: $0</b>"]
    D -- Yes --> F["Route to Premium Model<br/><b>Cost: Standard API Rate</b>"]

How it saves you money

Strategy	Savings	How it works
💾 Semantic Caching	100% on cache hits	Duplicate or semantically identical prompts served instantly from local cache
🆓 Free Agent Routing	100% on routine tasks	Standard coding queries routed to free-tier models or local endpoints
🎯 Strategic Fallbacks	50–90%	Premium models reserved for complex reasoning; routine tasks use cheap providers
📉 Quota Maximization	Subscription ROI ↑	Drain your paid subscriptions fully before falling back to API-key providers

💰 Real-world cost comparison

Scenario	Without OmniCode	With OmniCode	Savings
100 routine code completions/day	~$2.50/day (paid API)	$0 (free agent)	$75/mo
50 repeated similar queries	~$1.25/day (paid API)	$0 (semantic cache)	$37/mo
Subscription quota overflow	New API charges	Auto-fallback to free	Variable

🛡️ OmniProxy Native Control Center

A comprehensive, built-in management workspace embedded directly in the editor. No browser dashboards, no external tools — everything is native.

Section	Description
🏠 Home	Global runtime status, provider overview, model sync state, and proxy health
🔌 Providers	Connect, test, and manage integrations for 169+ AI providers with live status cards
🎯 Combos	Configure multi-model routing chains and automated fallback strategies
🧪 Batch Testing	Evaluate and test prompts across multiple models simultaneously
💰 Costs	Detailed tracking of token usage and financial spend per provider and model
📊 Analytics	Deep dive into request patterns, latency metrics, and usage trends
💾 Cache	Advanced semantic and prompt cache controls for optimized performance
📏 Limits & Quotas	Set rate limits, monitor quotas, and enforce budget controls
🖼️ Media	Manage image generation and rich media assets

🏗️ Architecture & How It Works

OmniCode's architecture is designed for speed, security, and seamless integration:

flowchart TD
    User([User Prompt]) --> ChatUI[Native Chat Interface]
    ChatUI --> OmniProxy{OmniProxy Router}
    
    OmniProxy -->|Check Cache| Cache[(Semantic Cache)]
    Cache -- Cache Hit --> ChatUI
    
    OmniProxy -- Cache Miss --> Evaluator{Combo / Fallback Rules}
    Evaluator -->|Free Tier| Free[Free Agent / Models]
    Evaluator -->|Premium Tier| Premium[Premium API Providers]
    Evaluator -->|Local/Custom| Custom[Custom Endpoints]
    
    Free --> ChatUI
    Premium --> ChatUI
    Custom --> ChatUI

Request flow

User Interaction: You submit a prompt via the native Chat interface or AI agent in the editor.
Model Selection: The editor retrieves available models from chatLanguageModels.json, continuously synced by the OmniProxy extension.
Smart Routing: The request passes through the OmniRoute Extension Bridge, which evaluates active routing rules, combos, and fallbacks.
Provider Execution: The Embedded Runtime securely authenticates using locally stored credentials (never committed to code) and dispatches to the appropriate provider.
Response & Caching: The response streams back in real-time, with optional semantic caching applied to reduce costs and latency on future requests.

Source map

Area	Path	Description
📦 Product metadata	`product.json`	Application identifiers, naming, and configuration
🎨 Resources	`resources/`	Icon assets for macOS, Windows, Linux, and web
🛡️ OmniProxy Dashboard	`src/vs/workbench/contrib/chat/browser/omniProxyManagement/`	Native management dashboard source code
⚙️ OmniProxy Runtime	`omniproxy-runtime/`	Embedded Node.js backend for AI routing
🔗 Provider Registry	`omniproxy-runtime/open-sse/config/providerRegistry.ts`	Single source of truth for all 169+ providers
🧩 Extension Bridge	`extensions/omniroute/`	Connects workbench UI to OmniProxy runtime
🧠 Model Discovery	`src/vs/workbench/contrib/chat/common/languageModels.ts`	Core logic for model discovery and endpoint integration

🤖 Personalized Coding Agent

Tailor your AI assistant to your exact workflow. Through Combos configuration in OmniProxy, you define:

🎯 Custom system prompts — personalize the AI's behavior and tone
🧠 Specialized context boundaries — scope the AI to your project's architecture
🔀 Preferred model fallbacks — set your ideal model chain
🔄 Hot-swap the brain — switch from local open-source to state-of-the-art paid models without changing your editing habits

The result: a deeply personalized coding agent that adapts to your coding style, no matter which backend AI provider is currently active.

📸 Screenshots & Showcase

OmniProxy Dashboard — Home

The dashboard provides an at-a-glance view of your runtime status, connected providers, synced models, and proxy configuration.

OmniProxy Showcase

🚀 Quick Start

Requirements

Requirement	Version
🖥️ macOS / Linux / Windows	Latest stable
📦 Node.js	`22.x` LTS (recommended)
📦 npm	`10.x`

1) Clone & install

git clone https://github.com/cyborgateuk-arch/OmniCode.git
cd OmniCode
npm install

2) Build

npm run gulp compile
node build/next/index.ts bundle --out out --target desktop

3) Run

# macOS
open -na '.build/electron/OmniCode.app' --args '.'

# Linux
.build/electron/OmniCode --no-sandbox .

# Windows
.build\electron\OmniCode.exe .

4) Setup OmniProxy Runtime

The OmniProxy runtime handles all AI requests locally and securely. On first run:

# Navigate to the runtime directory
cd omniproxy-runtime

# Copy the example environment file
cp .env.example .env

# Generate security secrets
openssl rand -base64 48    # JWT secret
openssl rand -hex 32       # API key encryption secret

Enter your provider API keys or OAuth credentials in the .env file. The runtime automatically initializes when you open the OmniProxy dashboard in the editor.

5) Connect a FREE provider (no signup)

Open the OmniProxy Dashboard → Providers → Connect Kiro AI (free Claude) or Qoder AI (free Qwen/DeepSeek) → done.

6) Start coding! 🎉

All connected models appear instantly in the native model picker. Select a model and start chatting.

🔒 Private & Secure by Design

Your keys, your machine, your data. OmniCode is designed with a security-first architecture.


🏠 100% local	🔐 Encrypted at rest
Runs entirely on your hardware. No OmniCode cloud in the request path.	API keys & OAuth tokens sealed with AES-256-GCM.
🚫 Zero telemetry by default	🛡️ Hardened gateway
Your prompts go only to the providers you choose, nowhere else.	PII sanitization, prompt-injection guards, loopback-only routes.
📜 MIT licensed	🔒 No secrets in source
Fully open-source — audit every line, self-host forever.	API keys, OAuth tokens, and secrets strictly excluded from version control.

Security highlights:

Sensitive values managed at runtime via secure secret storage or local .env file
.env and generated artifacts (logs, databases) strictly ignored by version control
PII sanitization and prompt injection guards built into the proxy layer
Database encryption at rest supported

See SECURITY.md for the full security policy.

🖥️ Where OmniCode Runs — Anywhere

Platform	Install	Highlights
🍎 macOS	`npm install && npm run gulp compile`	Native `.app` bundle with system tray
🐧 Linux	Same build pipeline	AppImage / deb / rpm packages
🪟 Windows	Same build pipeline	Native `.exe` with installer
🐳 Docker	`docker build`	Headless server mode for remote development
💻 From source	`git clone && npm install`	Full development setup

🛠️ Tech Stack

Component	Technology
Runtime	Node.js 22.x LTS
Language	TypeScript 6.0 — 100% TypeScript across `src/` and `omniproxy-runtime/`
Framework	Electron 39.x + VS Code workbench architecture
OmniProxy Runtime	Next.js 16 + React 19 + Tailwind CSS 4
Database	better-sqlite3 (SQLite) + LowDB (JSON) — routing decisions, cache, analytics
Auth	OAuth 2.0 (PKCE, Device Code, SSO OIDC) + API Keys + Encrypted Secret Storage
Streaming	Server-Sent Events (SSE) for real-time model responses
Security	AES-256-GCM encryption, PII sanitization, prompt injection guards
Testing	Mocha + Playwright (unit, integration, E2E)
CI/CD	GitHub Actions (automated builds and releases)
License	MIT

📖 Documentation

📘 Getting Started

Document	Description
Full OmniCode Documentation	Architecture, detailed flows, and build instructions
Quick Start	6-step install → connect → code
Copilot Instructions	Project architecture, coding guidelines, and validation steps

🔧 Operations & Configuration

Document	Description
Security Policy	Vulnerability reporting and security practices
Contributing Guide	Development setup and contribution guidelines
License	MIT License

🧠 Key Source Files

File	Purpose
`product.json`	Product identifiers, naming, and application metadata
`omniproxy-runtime/.env.example`	Complete environment variable reference
`omniproxy-runtime/open-sse/config/providerRegistry.ts`	Single source of truth for all provider configurations
`src/vs/workbench/contrib/chat/browser/omniProxyManagement/`	OmniProxy dashboard implementation
`extensions/omniroute/`	Extension bridge connecting UI to runtime

🤝 Contributing

We welcome contributions! Whether it's bug reports, feature requests, or pull requests — every contribution matters.

How to contribute

Fork the repository
Create your feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

Development setup

# Clone and install
git clone https://github.com/cyborgateuk-arch/OmniCode.git
cd OmniCode
npm install

# Start development watch mode
npm run watch

# Run tests
npm run test-node

# Type-check
npm run compile-check-ts-native

See CONTRIBUTING.md for detailed guidelines.

🔍 GitHub SEO & Discoverability

Topics & Keywords

ai-ide ai-coding ai-assistant code-editor omnicode omniproxy multi-model ai-routing free-ai developer-tools vscode-fork typescript electron llm-routing model-management ai-proxy cost-optimization semantic-caching oauth-providers coding-agent

What makes OmniCode searchable

AI-native IDE — not just a plugin, but a full editor with AI built in
Multi-provider routing — the only IDE that routes across 169+ AI providers
Free coding agent — $0 to start, with free-tier providers and semantic caching
Built-in dashboard — no external tools needed for provider management
Open source — MIT licensed, fully auditable, community-driven

📄 License

MIT License — free forever, open source, no restrictions.

Built with ❤️ for developers who demand more from their AI tools.

⭐ Star us on GitHub • 🐛 Report a Bug • 💬 Discussions

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
.agents/skills/launch		.agents/skills/launch
.config		.config
.devcontainer		.devcontainer
.eslint-plugin-local		.eslint-plugin-local
.github		.github
.vscode		.vscode
build		build
cli		cli
docs		docs
extensions		extensions
omniproxy-runtime		omniproxy-runtime
remote		remote
resources		resources
scripts		scripts
src		src
test		test
.editorconfig		.editorconfig
.eslint-ignore		.eslint-ignore
.git-blame-ignore-revs		.git-blame-ignore-revs
.gitattributes		.gitattributes
.gitignore		.gitignore
.lsifrc.json		.lsifrc.json
.mailmap		.mailmap
.mention-bot		.mention-bot
.npmrc		.npmrc
.nvmrc		.nvmrc
.vscode-test.js		.vscode-test.js
AGENTS.md		AGENTS.md
CONTRIBUTING.md		CONTRIBUTING.md
CodeQL.yml		CodeQL.yml
LICENSE.txt		LICENSE.txt
README.md		README.md
SECURITY.md		SECURITY.md
ThirdPartyNotices.txt		ThirdPartyNotices.txt
cglicenses.json		cglicenses.json
cgmanifest.json		cgmanifest.json
eslint.config.js		eslint.config.js
gulpfile.mjs		gulpfile.mjs
package-lock.json		package-lock.json
package.json		package.json
product.json		product.json
tsfmt.json		tsfmt.json

Folders and files

Latest commit

History

Repository files navigation

OmniCode

💥 The Promise

🤔 Why OmniCode?

🎯 Combos — Multi-Model Routing

⚡ How it works

🔀 Routing strategies

🧱 Resilience is built in

🏆 What Sets OmniCode Apart

🔌 169+ AI Providers — Unified Model Picker

🔐 OAuth Providers (One-Click Connect)

🔑 API Key Providers

🆓 Free Providers — $0, No Card

🔗 Custom Endpoints

📊 Cost Savings & Free Coding Agent

How it saves you money

💰 Real-world cost comparison

🛡️ OmniProxy Native Control Center

🏗️ Architecture & How It Works

Request flow

Source map

🤖 Personalized Coding Agent

📸 Screenshots & Showcase

OmniProxy Dashboard — Home

OmniProxy Showcase

🚀 Quick Start

Requirements

1) Clone & install

2) Build

3) Run

4) Setup OmniProxy Runtime

5) Connect a FREE provider (no signup)

6) Start coding! 🎉

🔒 Private & Secure by Design

🖥️ Where OmniCode Runs — Anywhere

🛠️ Tech Stack

📖 Documentation

📘 Getting Started

🔧 Operations & Configuration

🧠 Key Source Files

🤝 Contributing

How to contribute

Development setup

🔍 GitHub SEO & Discoverability

Topics & Keywords

What makes OmniCode searchable

📄 License

About

Topics

Resources

License

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases 1

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages