Skip to main content

Documentation Index

Fetch the complete documentation index at: https://gomodel.enterpilot.io/docs/llms.txt

Use this file to discover all available pages before exploring further.

GoModel routes Bedrock requests through the Bedrock Runtime Converse and ConverseStream APIs, normalizing responses across model families (Anthropic, Amazon Nova, Meta Llama, Mistral, Cohere, AI21, and others) into OpenAI-compatible chat completions. Bedrock has no API key of its own — GoModel resolves AWS credentials at runtime through the standard AWS credential chain (env, AWS_PROFILE, IAM Identity Center, instance/container roles).

​
Configure

BEDROCK_BASE_URL=us-east-1   # region (recommended) or full Bedrock endpoint URL
AWS_ACCESS_KEY_ID=...
AWS_SECRET_ACCESS_KEY=...
# AWS_SESSION_TOKEN=...      # only for temporary credentials
GOMODEL_MASTER_KEY=change-me
In production, prefer instance roles, IRSA, or IAM Identity Center over long-lived access keys. When BEDROCK_BASE_URL is empty, GoModel falls back to AWS_REGION / AWS_DEFAULT_REGION.

​
Run GoModel

docker run --rm -p 8080:8080 --env-file .env enterpilot/gomodel

​
Verify

curl -s http://localhost:8080/v1/chat/completions \
  -H "Authorization: Bearer $GOMODEL_MASTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic.claude-3-5-haiku-20241022-v1:0",
    "messages": [{"role": "user", "content": "Reply with the single word ok."}],
    "max_tokens": 80
  }'
GET /v1/models returns Bedrock models with owned_by set to the originating vendor (e.g. anthropic, amazon). Streaming (stream: true) and the Responses API are bridged onto Converse internally.

​
Restrict or extend the model list

By default GoModel queries Bedrock’s control plane and lists all on-demand text models the account has access to. To pin inference profile IDs or custom model ARNs: 
BEDROCK_MODELS=anthropic.claude-3-5-haiku-20241022-v1:0,amazon.nova-lite-v1:0
Or in config.yaml: 
providers:
  bedrock:
    type: bedrock
    base_url: "us-east-1"
    models:
      - anthropic.claude-3-5-haiku-20241022-v1:0
      - amazon.nova-lite-v1:0
Behavior follows the global CONFIGURED_PROVIDER_MODELS_MODE — fallback (default) uses the list only when ListFoundationModels is unavailable or empty; allowlist exposes only configured models and skips the upstream call.

​
Multiple Bedrock providers

Use suffixed env vars to register separate Bedrock instances: 
BEDROCK_US_BASE_URL=us-east-1
BEDROCK_US_MODELS=anthropic.claude-3-5-haiku-20241022-v1:0

BEDROCK_EU_BASE_URL=eu-west-1
BEDROCK_EU_MODELS=amazon.nova-lite-v1:0
This registers providers bedrock-us and bedrock-eu. Bedrock credentials still come from the standard AWS credential chain.

​
Not yet integrated

  • Bedrock embeddings — the InvokeModel embedding path is model-specific and not wired up.

​
Troubleshooting

  • AccessDeniedException / 403 — the AWS principal lacks Bedrock permissions, or model access has not been granted in the Bedrock console.
  • on-demand throughput isn't supported — the model requires an inference profile or provisioned throughput. Set BEDROCK_MODELS to the inference profile ID.
  • model registry has no models — region has no on-demand text models, or the control-plane call failed. Set BEDROCK_MODELS or check IAM.
  • Wrong region inferred — pass BEDROCK_BASE_URL explicitly. With a full endpoint URL, GoModel only extracts the region segment when the host ends in .amazonaws.com; custom endpoints should pair with AWS_REGION.

​
References