TeaQL Forge Rust Server

🇺🇸 English | 🇨🇳 中文

TeaQL Forge Server is a local-first code generation service for TeaQL, written entirely in Rust.

It turns TeaQL .xml domain models into Rust libraries and fully configured workspace scaffolds. It provides a highly concurrent axum-based backend to process code generation, validate models, and serve AI-assist prompts for modern agentic workflows (e.g., Cursor, Claude, etc.).

Features

• In-Memory Generation: Renders Rust files directly in memory and packages them as a streamable .zip file for lightning-fast execution.
• AI Assist API: Provides dynamic markdown templates (rust-assist-*) containing exact, model-aware code snippets to guide AI coding assistants in writing TeaQL standard code (CRUD, Queries, UI list pages, etc.).
• Live Previews & Dashboard: Serves an interactive HTML dashboard, a React Flow Entity Graph, and a Data Dictionary directly from the loaded model.
• Multiple Targets: Generates pure domain libraries (rust-lib-core) or fully configured CLI/Server workspaces (rust-app-console).
• Tiny Multi-Arch Docker: Built statically from scratch using cargo-zigbuild and musl. The final image is extremely lightweight (~5MB) and native to both linux/amd64 and linux/arm64.

Quick Start via Docker

The easiest way to run the local server is via Docker. This image runs smoothly on Linux, Windows (WSL), and Apple Silicon (M1/M2/M3) Macs.

docker run -d --name teaql-forge-server -p 8080:8080 teaql/teaql-forge-rs:latest

Once running, open your browser to view the interactive dashboard: 👉 http://localhost:8080/

Integrate with AI Assistants

Point your CLI tool or AI IDE to the local server to unlock real-time, model-aware code assistance:

export TEAQL_ENDPOINT_PREFIX=http://localhost:8080/

Usage with cargo-teaql

You can seamlessly integrate the local server with your existing cargo-teaql CLI workflow by pointing it to your local endpoint:

# Generate a full Workspace
cargo-teaql gen-workspace --endpoint-prefix http://127.0.0.1:8080/ models/model.xml

# Generate via the standard generate command
cargo teaql --input model.xml rust-app-console

Available API Endpoints

• GET / - Interactive HTML Dashboard with Live Previews.
• GET /version - Returns the current server version.
• GET /model-view.html - Interactive Entity Relationship Graph (React Flow).
• GET /data-design-react.html - Data Dictionary UI.
• GET /rust-assist-*/[entity] - Generates AI context prompts for specific operations (e.g., rust-assist-query/platform).
• POST /generate - Accepts multipart/form-data with an xml file and a scope (rust-lib-core or rust-app-console).
• POST /evaluate - Evaluates KSML scripts and expressions.

Building and Releasing

The server uses cargo-zigbuild for true multi-arch static compilation.

Local Development

cargo run --bin teaql-forge-server -- --host 127.0.0.1 --port 8080

Building Docker Images (Publishing)

To publish a new version of the Docker image to Docker Hub, use the automated script. This requires cargo-zigbuild installed locally (cargo install cargo-zigbuild).

# Publish as "latest"
./publish_docker.sh

# Publish a specific version tag
./publish_docker.sh v0.9.0

The script automatically:

1. Cross-compiles statically for x86_64 and aarch64.
2. Packages both binaries into scratch containers (-amd64 and -arm64).
3. Creates and pushes a unified Docker Manifest.

Configuration

When running the binary manually, you can pass arguments:

• --host: The interface to bind to (e.g., 0.0.0.0 to expose to the network, 127.0.0.1 for local-only).
• -p, --port: The port to listen on (default 8081).

Warning: Binding to 0.0.0.0 exposes the TeaQL local server to the network. Use Enterprise Mode or configure TLS/auth for production environments.

License

Apache-2.0 License