Bring your own key
Paste a key in Settings → Models & Providers. It is stored encrypted in your OS keychain and used only with that provider.
Models & providers
Managed models first.
Your own providers too.
Use the models included with Codvian credits, or connect a provider you already pay for.
CODVIAN Subscription Models
Included with a CODVIAN subscription. No provider keys needed.
Checking current availability…
- GPT-OSS 120B
- Qwen 3.8 Flash
- MiMo V2.6 Flash
- GLM-5.3 Flash
- DeepSeek V4.1 Flash
- MiniMax M3
- MiMo V2.6 Pro
- DeepSeek V4 Pro
- Qwen 3.8 Max
Displayed models may change over time and are subject to availability.
BYOK — Bring Your Own Key
73 supported providers
Use your own API key from the provider and pay the provider directly. BYOK usage is not part of CODVIAN credits.
- 302.AI
- Abacus
- abliteration.ai
- above.dev
- Agnes AI
- AI-ROUTER
- AIHubMix
- Aixy
- AKI.IO
- Alibaba
- Alibaba China
- Amazon Bedrock
- AMD
- Anthropic
- AnyAPI
- Arcee
- Azure OpenAI
- Baseten
- Cerebras
- Chutes
- Clarifai
- Cloudflare AI Gateway
- Cloudflare Workers AI
- Cohere
- Crusoe
- Databricks
- Deep Infra
- DeepSeek
- DigitalOcean
- Eden AI
- Fireworks AI
- Friendli
- GitHub Copilot
- GitLab Duo
- Groq
- Helicone
- Hugging Face
- Inference
- IO.NET
- Llama
- LM Studio
- MiniMax
- Mistral
- ModelScope
- Moonshot AI
- Morph
- NanoGPT
- Nebius
- NovitaAI
- NVIDIA
- Ollama
- OpenAI
- OpenRouter
- Perplexity
- Poe
- Poolside
- Qwen / Alibaba
- Requesty
- RunPod
- Sarvam AI
- Scaleway
- SiliconFlow
- Together AI
- Upstage
- v0
- Vercel AI Gateway
- Vertex AI
- watsonx.ai
- xAI
- Xiaomi
- Z.AI
- Zhipu AI
No provider matches. Any OpenAI-compatible endpoint can still be added as a custom provider.
Add any endpoint
Use “Add custom OpenAI-compatible provider” for gateways and hosts not listed here.
Switch any time
Change model or effort level for any message, without leaving the session.