Implement Anthropic Messages API support

[Copilot]

Adds Anthropic Messages API compatibility layer, enabling llama-server's native Anthropic support (merged in llama.cpp #17570) to be accessible through model-runner.

Changes

• New pkg/anthropic handler: Proxies requests to scheduler → llama.cpp backend

  • POST /anthropic/v1/messages - Chat completions with streaming
  • POST /anthropic/v1/messages/count_tokens - Token counting
  • Anthropic-format error responses

• Scheduler routing: Added Anthropic routes to routeHandlers(), recognized /v1/messages paths in backendModeForRequest()

• Request tracking: Added OriginAnthropicMessages origin constant for metrics

Usage

curl http://localhost:8080/anthropic/v1/messages \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ai/qwen3-coder",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Architecture mirrors the existing Ollama compatibility layer—thin proxy that forwards to llama.cpp which handles format conversion internally.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• docs.anthropic.com
  • Triggering command: /usr/bin/curl curl -s REDACTED (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Implement and test Anthropic Messages API</issue_title>
<issue_description>Qwen 3 Coder on Docker Hub would be a good model to test this with

ggml-org/llama.cpp#17570

The max context size an 36GB VRAM macbook pro can handle is:

llama-server -hf unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF:Q4_K_M -c 65536</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Implement and test Anthropic Messages API docker/model-runner#475

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.