oriveo
Open-source, bring-your-own-key (BYOK) multi-model AI chat client for iOS, Android and the web. One LLM client for OpenAI, Anthropic, Google Gemini, OpenRouter,
What is oriveo?
Open-source, bring-your-own-key (BYOK) multi-model AI chat client for iOS, Android and the web. One LLM client for OpenAI, Anthropic, Google Gemini, OpenRouter, DeepSeek and ten more providers, plus any OpenAI-, Anthropic- or Gemini-compatible endpoint — Ollama, LM Studio, llama.cpp, vLLM. Local-first, self-hostable, no account. AGPL-3.0.
Overview of oriveo
What it is
Oriveo Community Edition is an open-source, bring-your-own-key (BYOK) multi-model AI chat client for iOS, Android, and the web, with a native macOS client in development. It is a local-first alternative to hosted ChatGPT or Claude plans. Users supply their own API keys, and the client communicates directly with providers. No Oriveo account is required, and nothing reports back to Oriveo. It supports 15 model providers natively (OpenAI, Anthropic, Google Gemini, OpenRouter, DeepSeek, Grok, Mistral, Groq, Together AI, Fireworks AI, MiniMax, Z.ai, Qwen, Kimi (Moonshot), SiliconFlow) plus any OpenAI-, Anthropic- or Gemini-compatible endpoint, including llama.cpp, Ollama, LM Studio, vLLM running on the user's own machine.
Key features
- Chat with streaming, reasoning blocks, citations, attachments (images, video, PDF, Office documents, OpenDocument, EPUB, RTF, HTML, plain-text, source files), quote-a-selection, retry, regenerate, continue after interruption.
- 15 built-in providers with per-provider model and generation-parameter overrides, and choice of regional endpoint.
- Relay: any OpenAI-, Anthropic- or Gemini-compatible endpoint, plus llama.cpp's native API, including on LAN.
- Local model servers: llama.cpp, Ollama, LM Studio, vLLM, Open WebUI; iOS and Android discover them via mDNS or port probing.
- Subscription sign-in: use ChatGPT or Grok subscription instead of API key via device-authorization flow.
- Skills: reusable system prompts with own model, reasoning setting, and reference documents.
- MCP tools: add remote MCP server, sign in, let model call tools in chat; each tool configurable as own, ask-first, or off.
- Notes and folders: capture replies as notes, organize conversations, search across both.
- Cross-check: hand an answer to a second model for review.
- Memory: store a few facts about yourself, carried into every new conversation.
- Cost tracking: per-message and per-provider spend computed on device from actual response data, including cache tiers.
- Image generation where provider supports it.
- Backup: export everything to a file; provider keys optionally encrypted with a password.
- 16 interface languages, including full right-to-left layout for Arabic.
Use cases
- Users who want to use multiple AI models from different providers in a single interface without paying a subscription to a middleman.
- Users who already have API keys for providers like OpenAI, Anthropic, Google Gemini, etc., and want a local-first client that does not send data to a third-party service.
- Self-hosting the web client on one's own machine for full control.
- Running local model servers (Ollama, LM Studio, etc.) and accessing them from mobile or web clients.
- Using subscription-based access (ChatGPT, Grok) without an API key.
Getting started
See the repository README for build instructions.
License and pricing
AGPL-3.0 open source.
Based on the project's public README and repository metadata.