modelmux 1.0.0

ModelMux - high-performance Rust gateway that translates OpenAI-compatible API requests to Vertex AI (Claude), with streaming, tool calling, and production-grade reliability.
Documentation
# ModelMux Development Task Tracker

> **Purpose**: Single entry point for current development work. Links to detailed task files in `tasks/`.  
> **Source**: [plan.md]plan.md, [AGENT.md]AGENT.md

## Current Focus

**Phase 1: Foundation & Quick Wins** (Weeks 1–3)

### Recommended Order to Start

1. **[TASK-004]tasks/TASK-004-openai-compatible-provider.md** — OpenAI-compatible provider (company priority for vLLM)
2. **[TASK-005]tasks/TASK-005-provider-configuration.md** — Provider config in TOML (depends on TASK-004)
3. **[TASK-001]tasks/TASK-001-integration-tests.md** — Integration tests (can run in parallel)
4. **[TASK-007]tasks/TASK-007-docker-image.md** — Docker image

### Quick Commands

```bash
# Run tests
cargo test

# Run with debug logging
RUST_LOG=debug cargo run

# Validate config
modelmux config validate
```

## Task Index

**Phase 1 (Weeks 1–3):** TASK-001 through TASK-007  
**Phase 2 (Weeks 4–7):** TASK-008 through TASK-012  
**Phase 3 (Weeks 8–11):** TASK-013 through TASK-022  
**Phase 4 (Ongoing):** TASK-023+

| # | Task | Description |
|---|------|-------------|
| 001 | [Integration tests]tasks/TASK-001-integration-tests.md | Tests, fixtures, converter tests |
| 002 | [Error handling]tasks/TASK-002-error-handling.md | Error variants, request ID, structured logging |
| 003 | [Documentation pass]tasks/TASK-003-documentation-pass.md | API_COMPATIBILITY, TROUBLESHOOTING, README |
| 004 | [OpenAI-compatible provider]tasks/TASK-004-openai-compatible-provider.md | vLLM, passthrough, auth |
| 005 | [Provider configuration]tasks/TASK-005-provider-configuration.md | TOML schema, factory, presets |
| 006 | [vLLM integration]tasks/TASK-006-vllm-integration-testing.md | Test script, VLLM_SETUP, failover |
| 007 | [Docker & releases]tasks/TASK-007-docker-image.md | Dockerfile, Compose, GitHub Actions |
| 008–012 | [Responses API]tasks/TASK-008-responses-api-research.md | Research, types, converters, endpoint |
| 013–022 | [Polish]tasks/TASK-013-prometheus-metrics.md | Metrics, logging, caching, docs |
| 023 | [Contributor guidelines]tasks/TASK-023-contributor-guidelines.md | CONTRIBUTING, ARCHITECTURE |

Full index with plan step mapping: [tasks/README.md](tasks/README.md)

## How to Use (per AGENT.md)

1. **Pick a task** — Open the task file
2. **Update phase status**`pending``in_progress``complete`
3. **Log errors** — Add to "Errors Encountered" in the task file
4. **Re-read plan** before major decisions
5. **Never repeat failed actions** — Mutate approach after failures

## Next Actions (from plan.md)

- [ ] Create `tests/integration/` directory
- [ ] Write first integration test
- [ ] Start `OpenAiCompatibleProvider` implementation in `src/provider.rs`
- [ ] Update config schema for provider selection