Gateways

A gateway is the API service between taoCode and model inference. Your gateway determines authentication, available models, pricing, rate limits, and data-handling terms.

Chutes · Subnet 64

Setting	Value
API base	https://llm.chutes.ai/v1
API key variable	CHUTES_API_KEY
Stored credential	tao-chutes
Account	Get a Chutes key

export CHUTES_API_KEY="your-chutes-key"
taocode models

Targon · Subnet 4

Setting	Value
API host	https://api.targon.com
API key variable	TARGON_API_KEY
Stored credential	tao-targon
Account	Visit Targon

export TARGON_API_KEY="your-targon-key"
taocode models

Connect more than one gateway

Set each gateway’s key or use /connect in the interactive session. Select a model using its full tao/<gateway>/<model> reference. Credentials and billing remain specific to each gateway.

Check capabilities before a task

A coding workflow needs a model and endpoint that support tool calls. Context limits, availability, and confidential-compute support can differ between models on the same gateway.

A gateway capability label is not proof that a particular response was verified. Read Verifiable inference and consult the gateway’s current documentation for the endpoint you use.

Troubleshoot a connection

Check for a revoked key, an empty account balance, a gateway outage, or a stale stored credential. Reconnect before retrying, then choose a reference returned by taocode models.