Amp

Configuration

• Create a CLIProxyAPI config.yaml with a prismatic custom provider.
• Set provider base-url to https://api.prismaticapi.com/v1 and api-key to your Prismatic key.
• Map Amp requested model names to Prismatic public model IDs under model-mappings.
• Start the local proxy on http://localhost:3456, then point Amp ANTHROPIC_BASE_URL to that proxy.

Example

server:
  port: 3456

custom-providers:
  - name: "prismatic"
    base-url: "https://api.prismaticapi.com/v1"
    api-key: "pa_YOUR_KEY"
    models:
      - "claude-sonnet-4-6"
      - "claude-haiku-4-5"

model-mappings:
  gpt-5.4:
    provider: prismatic
    model: claude-sonnet-4-6
  gpt-5.3-codex:
    provider: prismatic
    model: claude-sonnet-4-6
  glm-5-turbo:
    provider: prismatic
    model: claude-haiku-4-5

force-model-mappings: true

docker run -d --name cliproxyapi \
  -p 3456:3456 \
  -v ./config.yaml:/CLIProxyAPI/config.yaml \
  eceasy/cliproxyapi

export ANTHROPIC_BASE_URL="http://localhost:3456"
export ANTHROPIC_API_KEY="pa_YOUR_KEY"

amp

Recommended models

• claude-sonnet-4-6 — strong model for reasoning, refactors and coding agents.
• claude-haiku-4-5 — fast option for autocomplete, small questions and lightweight tasks.
• deepseek-v3-2 or another public ID — alternative based on your active catalog and plan.

Troubleshooting

Notes

• Amp calls the local proxy in this setup; CLIProxyAPI translates the request to the Prismatic OpenAI-compatible API.
• Keep force-model-mappings: true so unexpected Amp model names are mapped deliberately.
• Use public model IDs returned by GET /v1/models when you customize model-mappings.