Configuration¶

Environment Variables¶

Variable	CLI Flag	Description	Default
`OPENAI_API_KEY`	`--api-key`	API key for authentication	required
`OPENAI_BASE_URL`	`--base-url`	Base URL for OpenAI-compatible API	required
`HOST`	`--host`	IP/hostname for the API server	`localhost`
`CAPABILITIES`	`--capabilities`, `-c`	Extra capabilities to mark the model as supporting	`[]`
`MODELS`	`--models`, `-m`	Extra models to include in `/api/tags`	`[]`

Capabilities¶

Capabilities indicate what features a model supports. Available capabilities:

tools - Function calling / tools support
insert - Insert mode
vision - Vision / image input support
embedding - Embedding generation
thinking - Thinking mode
completion - Always included, basic completion

Setting Capabilities¶

# Single capability
oai2ollama -c tools

# Multiple capabilities (separate flags)
oai2ollama -c tools -c vision

# Multiple capabilities (comma-separated)
oai2ollama --capabilities tools,vision

Extra Models¶

Add models that aren't returned by the upstream API:

# Multiple models (separate flags)
oai2ollama -m model1 -m model2

# Multiple models (comma-separated)
oai2ollama --models model1,model2

Example .env File¶

OPENAI_API_KEY=your_api_key
OPENAI_BASE_URL=https://api.openai.com/v1
HOST=0.0.0.0
CAPABILITIES=["vision","thinking"]
MODELS=["custom-model1","custom-model2"]

Docker Example¶

docker run -p 11434:11434 \
  -e OPENAI_API_KEY="your_api_key" \
  -e OPENAI_BASE_URL="your_base_url" \
  -e CAPABILITIES='["vision","thinking"]' \
  oai2ollama

Or with CLI arguments:

docker run -p 11434:11434 \
  oai2ollama --api-key your_api_key --base-url your_base_url \
  --capabilities tools,vision --models custom-model