OpenAI Compatibility

Parel is designed as a drop-in replacement for OpenAI-compatible chat, image, audio, and embedding calls. In most applications, the migration is two values:

client = OpenAI(
    api_key=os.environ["OPENAI_API_KEY"],
    base_url="https://api.openai.com/v1",
    api_key=os.environ["PAREL_API_KEY"],
    base_url="https://api.parel.cloud/v1",
)

Compatible surfaces

Use case	Endpoint
Chat and vision	`POST /v1/chat/completions`
Legacy completions / fill-in-the-middle	`POST /v1/completions`
Image generation	`POST /v1/images/generations`
Image editing	`POST /v1/images/edits`
Speech to text	`POST /v1/audio/transcriptions`
Text to speech	`POST /v1/audio/speech`
Embeddings	`POST /v1/embeddings`
Models	`GET /v1/models`

Parel extensions

Parel also adds endpoints for model comparison, imported models, dedicated GPU deployments, usage, and billing. These are documented separately and use the same Authorization: Bearer header.

Response headers

Most paid API responses include useful telemetry:

Header	Meaning
`X-Parel-Cost-USD`	Request cost
`X-Parel-Budget-Remaining`	Remaining account balance
`X-Parel-Latency-Ms`	Gateway latency
`X-Parel-Provider`	Selected upstream provider when available