How the connection works
Open WebUI sends chat requests to the URL of each OpenAI connection. FreeLLMAPI listens on http://localhost:3001/v1 and accepts the same requests. It selects a free model, sends the request and returns the answer in OpenAI format.
- One connection, many providers: you add one URL and one unified key. Your provider keys stay in FreeLLMAPI on your computer.
- Failover: when a provider returns a 429 error, the router tries the next model in your fallback chain.
- Limits: free tiers have request and token limits. More provider keys give more requests per day. Failover spreads the load, but it does not remove the limits.
Install FreeLLMAPI first. It runs in Docker. Open the dashboard at http://localhost:3001, add at least two free provider keys on the Keys page and copy your unified key.
Install FreeLLMAPI (macOS or Linux)
curl -fsSL https://freellmapi.co/install.sh | bashStep 1: Find the correct URL
The URL that Open WebUI must use depends on where each program runs. Open WebUI usually runs in Docker. Inside a container, localhost is the container itself, not your computer.
| Open WebUI runs | FreeLLMAPI runs | URL to use |
|---|---|---|
| On the host (pip or uv) | On the host (Docker or desktop app) | http://localhost:3001/v1 |
| In Docker Desktop (macOS, Windows) | On the host | http://host.docker.internal:3001/v1 |
| In Docker on Linux | In Docker (install script) | http://freellmapi:3001/v1 on a shared Docker network |
| On another computer | On a server on your LAN | http://<server-ip>:3001/v1, with HOST_BIND=0.0.0.0 on a trusted network only |
By default, the FreeLLMAPI container publishes its port on 127.0.0.1 only. On Linux, a second container cannot always reach that port through the host. Connect both containers to one Docker network. Then use the service name freellmapi as the host name.
Connect Open WebUI to the FreeLLMAPI network (Linux)
# Find the network of the FreeLLMAPI stack (usually freellmapi_default)
docker network ls
# Attach the Open WebUI container to it
docker network connect freellmapi_default open-webui
# Test the URL from inside the Open WebUI container
docker exec open-webui curl -s http://freellmapi:3001/v1/models \
-H "Authorization: Bearer <unified-key>" | head -c 300Step 2: Add the connection
You can add the connection in the interface or with environment variables. The interface is the simplest method.
- Open the Admin Panel and go to Settings > Connections.
- Find the Manage OpenAI API Connections list. Click Add Connection.
- In URL, type the URL from Step 1, for example
http://host.docker.internal:3001/v1. - In API Key, paste your FreeLLMAPI unified key.
- Click Verify Connection. Then click Save.
To set the connection when the container starts, use OPENAI_API_BASE_URL and OPENAI_API_KEY. The command below is the standard Open WebUI docker run command with these two variables added.
docker run with the FreeLLMAPI connection
docker run -d -p 3000:8080 \
--add-host=host.docker.internal:host-gateway \
-e OPENAI_API_BASE_URL=http://host.docker.internal:3001/v1 \
-e OPENAI_API_KEY=<unified-key> \
-v open-webui:/app/backend/data \
--name open-webui --restart always \
ghcr.io/open-webui/open-webui:mainTo keep an existing OpenAI connection, use the plural variables. OPENAI_API_BASE_URLS and OPENAI_API_KEYS take lists separated by semicolons. The first key goes with the first URL.
Two connections in one container
-e OPENAI_API_BASE_URLS="https://api.openai.com/v1;http://host.docker.internal:3001/v1" \
-e OPENAI_API_KEYS="<openai-key>;<unified-key>"ENABLE_PERSISTENT_CONFIG=False to make the environment win.Step 3: Keep the model list short
FreeLLMAPI returns its full catalog on /v1/models. The catalog also contains models that you cannot call yet, because you did not add a key for them. Open WebUI shows every model that the connection returns. A long list of models that fail is not useful.
Open the connection again and use the Model IDs field. When you add IDs, Open WebUI shows only these IDs and stops reading the full list. Add the router entries first:
| Model ID | What the router does |
|---|---|
auto | Follows your active fallback chain from the dashboard |
auto:fast | Ranks enabled models by measured speed |
auto:smart | Ranks enabled models by intelligence |
auto:reliable | Ranks enabled models by recent success rate |
auto:<profile> | Uses a named fallback chain, for example a vision chain |
You can also add a fixed model ID, for example a model from the table below. A fixed model gives the same style of answer every time. It also stops when that model's free limit is used, unless the same model is free on other providers.
To save quota, set the task model to a fast entry. Open WebUI uses the task model for titles and search queries. The variable is TASK_MODEL_EXTERNAL, for example TASK_MODEL_EXTERNAL=auto:fast.
| Model | Free on | Context | Free limits (best offer) | Tools | Vision |
|---|---|---|---|---|---|
| GLM 5.3 | APINex, UnoRouter | 1M | 5 req/min 10M weighted tokens/day shared across APINex free routes; 4x token multiplier; 5 requests/minute/account | Yes | No |
| GLM 5.3 Flash | APINex, BlazeAPI, NVIDIA NIM +1 more | 1M | 40 req/min 40 RPM | Yes | Yes |
| Gemini 3.8 Flash | APINex, Google AI Studio | 1M | 5 req/min 10M weighted tokens/day shared across APINex free routes; 3x token multiplier; 5 requests/minute/account | Yes | Yes |
| Qwen3.8 Flash Next | AMD Radeon Cloud, BlazeAPI, ElectronHub | 262K | Recurring daily account allowance | Yes | No |
| Qwen3.8 Flash | ElectronHub | 1M | $0.25/week shared recurring credit | Yes | Yes |
| Deepseek V4.1 Flash | APINex | 1M | 5 req/min 10M weighted tokens/day shared across APINex free routes; 3x token multiplier; 5 requests/minute/account | No | No |
| Gemini 3.7 Flash | Google AI Studio | 1M | 10 req/min 20 req/day 250,000 tokens/min ~3M | Yes | Yes |
| GPT-5.6 Luna | ElectronHub | 1M | $0.25/week shared recurring credit | Yes | Yes |
| Dots3 Note Preview | Kilo Gateway, OpenRouter | 512K | 20 req/min 50 req/day 50 req/day account-wide (1,000 after a $10 purchase) | Yes | Yes |
| DeepSeek V4 Flash | AMD Radeon Cloud, BazaarLink, ElectronHub | 1M | 10 req/min 50 req/day Shared rate limit | Yes | No |
Showing 10 of 184 matching free models. See all models.
Live data from the FreeLLMAPI catalog (version 2026.11.19, checked 19 November 2026). Free limits change often. The catalog is checked every day.
If you send images in Open WebUI, the router uses only vision models for that request. If no vision model is in your chain, the request returns error 422. These free models accept images:
| Model | Free on | Context | Free limits (best offer) | Tools | Vision |
|---|---|---|---|---|---|
| GLM 5.3 Flash | UnoRouter | 1M | 1 req/min 1 req/min per model | Yes | Yes |
| Gemini 3.8 Flash | Google AI Studio | 1M | Recurring provider quota | Yes | Yes |
| Gemini Flash (latest alias) | Google AI Studio | - | 10 req/min 20 req/day 250,000 tokens/min ~3M | Yes | Yes |
| Qwen3.8 Flash | ElectronHub | 1M | $0.25/week shared recurring credit | Yes | Yes |
| Gemini 3.7 Flash | Google AI Studio | 1M | 10 req/min 20 req/day 250,000 tokens/min ~3M | Yes | Yes |
| GPT-5.6 Luna | ElectronHub | 1M | $0.25/week shared recurring credit | Yes | Yes |
Showing 6 of 67 matching free models. See all models.
Live data from the FreeLLMAPI catalog (version 2026.11.19, checked 19 November 2026). Free limits change often. The catalog is checked every day.
Use free embeddings for documents
Open WebUI can make embeddings for your documents through an OpenAI-compatible API. FreeLLMAPI serves /v1/embeddings. Embedding failover never changes the model, so the vectors in your store stay compatible.
Embedding settings for Open WebUI
-e RAG_EMBEDDING_ENGINE=openai \
-e RAG_OPENAI_API_BASE_URL=http://host.docker.internal:3001/v1 \
-e RAG_OPENAI_API_KEY=<unified-key> \
-e RAG_EMBEDDING_MODEL=bge-m3Select one embedding family and keep it. If you change the family, make the embeddings for all documents again. The free embedding models in today's catalog are:
| Model | Provider | Dimensions | Max input | Free limit |
|---|---|---|---|---|
Gemini Embeddinggemini-embedding-001 | Google AI Studio | 3072 | 2048 | 100 rpm · 1K req/day |
Nemotron Embed VL 1Bnvidia/llama-nemotron-embed-vl-1b-v2 | NVIDIA NIM | 2048 | 8192 | ~40 rpm |
Nemotron 3 Embed 1B (OR free)nvidia/nemotron-3-embed-1b:free | OpenRouter | 2048 | 32768 | 20 rpm · shared free daily cap |
SEA-LION E5 Embedding 600Maisingapore/SEA-LION-E5-Embedding-600M | SEA-LION | 1024 | 8192 | free · 10 rpm |
SEA-LION ModernBERT Embedding 300Maisingapore/SEA-LION-ModernBERT-Embedding-300M | SEA-LION | 1024 | 8192 | free · 10 rpm |
SEA-LION ModernBERT Embedding 600Maisingapore/SEA-LION-ModernBERT-Embedding-600M | SEA-LION | 1024 | 8192 | free · 10 rpm |
BGE-M3@cf/baai/bge-m3 | Cloudflare Workers AI | 1024 | 8192 | 10K neurons/day (shared) |
EmbeddingGemma 300M@cf/google/embeddinggemma-300m | Cloudflare Workers AI | 768 | 2048 | 10K neurons/day (shared) |
Qwen3 Embedding 0.6B@cf/qwen/qwen3-embedding-0.6b | Cloudflare Workers AI | 1024 | 8192 | 10K neurons/day (shared) |
Cohere Embed v4embed-v4.0 | Cohere | 1536 | 128000 | 1K calls/mo (shared w/ chat) |
Gemini Embedding 2gemini-embedding-2 | Google AI Studio | 3072 | 8192 | free-tier rate limited |
Nemotron 3 Embed 1Bnvidia/nemotron-3-embed-1b | NVIDIA NIM | 2048 | 4096 | ~40 rpm |
PLaMo Embedding 1B (CF)@cf/pfnet/plamo-embedding-1b | Cloudflare Workers AI | 2048 | 4096 | 10K neurons/day (shared) |
BGE Small EN v1.5 (CF)@cf/baai/bge-small-en-v1.5 | Cloudflare Workers AI | 384 | 512 | 10K neurons/day (shared) |
BGE Base EN v1.5 (CF)@cf/baai/bge-base-en-v1.5 | Cloudflare Workers AI | 768 | 512 | 10K neurons/day (shared) |
BGE Large EN v1.5 (CF)@cf/baai/bge-large-en-v1.5 | Cloudflare Workers AI | 1024 | 512 | 10K neurons/day (shared) |
Mistral Embedmistral-embed | Mistral | 1024 | 8192 | free mode · monthly cap (shared) |
Codestral Embedcodestral-embed | Mistral | 1536 | 8192 | free mode · monthly cap (shared) |
Llama Text Embed v2 (Pinecone)llama-text-embed-v2 | Pinecone | 1024 | 2048 | 5M tokens/mo per model |
Multilingual E5 Large (Pinecone)multilingual-e5-large | Pinecone | 1024 | 507 | 5M tokens/mo per model |
Nemotron 3 Embed 1B (Speka)nvidia/nemotron-3-embed-1b | Speka | 2048 | 8192 | $1/month shared across Speka models |
OpenAI: Text Embedding 3 Small (ElectronHub)text-embedding-3-small | ElectronHub | 1536 | 8192 | $0.25/week shared free credits |
OpenAI: Text Embedding 3 Large (ElectronHub)text-embedding-3-large | ElectronHub | 3072 | 8192 | $0.25/week shared free credits |
Nemotron 3 Embed 8B (Inferbase)nvidia-nemotron-3-embed-8b | Inferbase | 4096 | 8192 | 5,000 requests/month; 100 requests and $0.50 usage/day shared |
Qwen 3 Embedding 0.6B (Inferbase)qwen3-embedding-0-6b | Inferbase | 1024 | 8192 | 5,000 requests/month; 100 requests and $0.50 usage/day shared |
Qwen 3 Embedding 4B (Inferbase)qwen3-embedding-4b | Inferbase | 2560 | 8192 | 5,000 requests/month; 100 requests and $0.50 usage/day shared |
Qwen 3 Embedding 8B (Inferbase)qwen3-embedding-8b | Inferbase | 4096 | 8192 | 5,000 requests/month; 100 requests and $0.50 usage/day shared |
LFM2.5 Embedding 350M (LLMTR)liquid/lfm-2.5-embedding-350m-free | LLMTR | 1024 | 512 | Ongoing zero-price route; provider terms apply |
BGE-M3 (SEA-LION)BAAI/bge-m3 | SEA-LION | 1024 | 8192 | free · 10 rpm |
Cohere Embed English v3embed-english-v3.0 | Cohere | 1024 | 512 | 1K calls/mo (shared w/ chat) |
Qwen3 Embedding 8B (OVH)Qwen3-Embedding-8B | OVH AI Endpoints | 4096 | 32768 | free · anonymous 2 req/min per IP |
BGE-M3 (OVH)bge-m3 | OVH AI Endpoints | 1024 | 8192 | free · anonymous 2 req/min per IP |
BGE Multilingual Gemma2 (OVH)bge-multilingual-gemma2 | OVH AI Endpoints | 3584 | 8192 | free · anonymous 2 req/min per IP |
Mistral: Codestral Embed (ElectronHub)codestral-embed | ElectronHub | 1536 | 8192 | $0.25/week shared free credits |
Mistral: Mistral Embed (ElectronHub)mistral-embed | ElectronHub | 1024 | 8192 | $0.25/week shared free credits |
Google: Gemini Embedding 2 (ElectronHub)gemini-embedding-2 | ElectronHub | 3072 | 8192 | $0.25/week shared free credits |
Google: Gemini Embedding 001 (ElectronHub)gemini-embedding-001 | ElectronHub | 3072 | 2048 | $0.25/week shared free credits |
OpenAI: Text Embedding Ada 002 (ElectronHub)text-embedding-ada-002 | ElectronHub | 1536 | 8192 | $0.25/week shared free credits |
Cohere Embed English Light v3embed-english-light-v3.0 | Cohere | 384 | 512 | 1K calls/mo (shared w/ chat) |
Cohere Embed Multilingual v3embed-multilingual-v3.0 | Cohere | 1024 | 512 | 1K calls/mo (shared w/ chat) |
Cohere Embed Multilingual Light v3embed-multilingual-light-v3.0 | Cohere | 384 | 512 | 1K calls/mo (shared w/ chat) |
Live data from the FreeLLMAPI catalog (version 2026.11.19, checked 19 November 2026). Free limits change often. The catalog is checked every day.
Troubleshooting
| Problem | Cause | Fix |
|---|---|---|
| Verify Connection fails with a network error | localhost points to the Open WebUI container | Use host.docker.internal or a shared Docker network, as in Step 1 |
| Verify Connection fails with 401 | The key is wrong or has spaces | Copy the unified key again from the Keys page |
| Hundreds of models in the selector | The connection reads the full catalog | Add auto and your selected IDs in Model IDs |
| A model fails with every message | No key for that model's provider | Add the provider key, or select auto |
| Error 429 after many chats | All provider keys in the chain are at their limit | Add keys from more providers, or wait for the reset time in the Retry-After header |
| Environment variables have no effect | Values saved in the database take precedence | Edit the connection in the Admin Panel |
To see which provider answered a chat, read the X-Routed-Via header of the response. The Server Logs viewer in the dashboard Analytics menu also shows routing decisions and key cooldowns.
Frequently asked questions
Is there a free API for Open WebUI?
Yes. Open WebUI works with any OpenAI-compatible API. FreeLLMAPI combines free provider tiers behind one URL and one key. Each free tier has its own limits.
What URL do I use for an OpenAI-compatible connection in Open WebUI?
Use the base URL with /v1 at the end, for example http://host.docker.internal:3001/v1. Do not add /chat/completions. If Open WebUI runs in Docker, do not use localhost.
Why does Open WebUI ignore my OPENAI_API_BASE_URL change?
Open WebUI saves connection settings in its database after the first start. Values in the database take precedence. Edit the connection in the Admin Panel, or set ENABLE_PERSISTENT_CONFIG=False.
Can Open WebUI use several free providers at once?
Yes. You can add one connection per provider. With FreeLLMAPI, one connection is enough, because the router selects the provider and moves to the next one when a limit stops a request.
Which free model is best in Open WebUI?
Start with auto. It follows your fallback chain. For a fixed model, select one with a large context window from the table on this page.
Related guides
- Free API for SillyTavern
- Free OpenAI-compatible API base URLs
- Free embeddings API
- Free LLM API for RAG
- Free LLM API rate limits
Sources
- Open WebUI docs: environment variable configuration (checked 20 November 2026)
- Open WebUI docs: connect an OpenAI-compatible provider (checked 20 November 2026)
- Open WebUI README: Docker installation (checked 20 November 2026)
- FreeLLMAPI docs: REST API (routing strategies, model listing, embeddings) (checked 20 November 2026)
- FreeLLMAPI docs: Docker deployment (checked 20 November 2026)