# Antigravity Proxy
OpenAI-compatible proxy that intercepts and relays requests to Google's Antigravity language server, impersonating the real Electron webview.
## Quick Start
```bash
# Build
cargo build --release
# Run (language server must be running)
RUST_LOG=info ./target/release/antigravity-proxy
# Custom port
RUST_LOG=info ./target/release/antigravity-proxy --port 9000
```
Default port: **8741**
## Endpoints
| Method | Path | Description |
| -------- | ---------------------- | ----------------------------------------------------------- |
| `POST` | `/v1/responses` | **Responses API** (primary) — supports `stream: true/false` |
| `POST` | `/v1/chat/completions` | Chat Completions API (OpenAI compat shim) |
| `GET` | `/v1/models` | List available models |
| `GET` | `/v1/sessions` | List active sessions |
| `DELETE` | `/v1/sessions/:id` | Delete a session |
| `POST` | `/v1/token` | Set OAuth token at runtime |
| `GET` | `/v1/usage` | MITM-intercepted token usage stats |
| `GET` | `/v1/quota` | LS quota — credits, per-model rate limits, reset timers |
| `GET` | `/health` | Health check |
## Available Models
| Name | Label |
| ------------------- | ---------------------------------------- |
| `opus-4.6` | Claude Opus 4.6 (Thinking) — **default** |
| `opus-4.5` | Claude Opus 4.5 (Thinking) |
| `gemini-3-pro-high` | Gemini 3 Pro (High) |
| `gemini-3-pro` | Gemini 3 Pro (Low) |
| `gemini-3-flash` | Gemini 3 Flash |
## Example: Responses API
### Sync
```bash
curl -s http://localhost:8741/v1/responses \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-3-flash",
"input": "Say hello in exactly 3 words",
"stream": false,
"timeout": 60
}' | jq .
```
### Streaming
```bash
curl -N http://localhost:8741/v1/responses \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-3-flash",
"input": "Say hello in exactly 3 words",
"stream": true,
"timeout": 60
}'
```
### Multi-turn (session reuse)
```bash
# First message
curl -s http://localhost:8741/v1/responses \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-3-flash",
"input": "What is 2+2?",
"conversation": "my-session-1",
"stream": false
}' | jq .
# Follow-up in same cascade
curl -s http://localhost:8741/v1/responses \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-3-flash",
"input": "Now multiply that by 10",
"conversation": "my-session-1",
"stream": false
}' | jq .
```
## Authentication
The proxy needs an OAuth token. Three ways to provide it:
1. **Environment variable**: `export ANTIGRAVITY_OAUTH_TOKEN=ya29.xxx`
2. **Token file**: `echo 'ya29.xxx' > ~/.config/antigravity-proxy-token`
3. **Runtime API**: `curl -X POST http://localhost:8741/v1/token -d '{"token":"ya29.xxx"}'`
## Stealth Features
- **TLS fingerprint**: BoringSSL with Chrome 142 JA3/JA4 + H2 fingerprint via `wreq`
- **Protobuf**: Hand-rolled encoder producing byte-exact match to real webview traffic
- **Warmup**: Mimics real webview startup RPC calls
- **Heartbeat**: Periodic keep-alive matching real webview lifecycle
- **Jitter**: Randomized polling intervals to avoid automation fingerprint
- **Session reuse**: Cascades are reused for multi-turn, matching real webview behavior
- **Version detection**: Auto-detects Antigravity/Chrome/Electron versions from installed app
## MITM Proxy
Built-in TLS-intercepting proxy captures real token usage from LS ↔ Google/Anthropic traffic. Disabled with `--no-mitm`.
### Setup
```bash
# 1. Start proxy (generates CA cert automatically)
RUST_LOG=info ./target/release/antigravity-proxy
# 2. Install wrapper (patches LS binary to route through MITM)
sudo ./scripts/mitm-wrapper.sh install
# 3. Restart Antigravity — done!
# Check status
./scripts/mitm-wrapper.sh status
# Uninstall
sudo ./scripts/mitm-wrapper.sh uninstall
```
### Usage Stats
```bash
curl -s http://localhost:8741/v1/usage | jq .
```
## Standalone Language Server
Launch an isolated LS instance for experimentation:
```bash
# Basic test (starts, checks quota, exits)
./scripts/standalone-ls.sh
# Foreground mode (stays alive)
./scripts/standalone-ls.sh --fg
# With MITM traffic interception
./scripts/standalone-ls.sh --mitm
# Capture a clean traffic snapshot
./scripts/standalone-ls.sh --snapshot
# Snapshot with custom prompt
./scripts/standalone-ls.sh --snapshot --prompt "Explain quantum computing"
```
The standalone LS shares the main Antigravity app's OAuth (via its extension server) but has its own port, data directory, and cascades.
### Traffic Snapshots
The `--snapshot` flag captures all HTTP/2 traffic and formats it into a clean, color-coded report:
```
══════════════════════════════════════════════════════════════════════
STANDALONE LS TRAFFIC SNAPSHOT
══════════════════════════════════════════════════════════════════════
▸ Outbound Connections
→ antigravity-unleash.goog (Feature Flags)
→ play.googleapis.com (Telemetry)
══════════════════════════════════════════════════════════════════════
antigravity-unleash.goog — Feature Flags
══════════════════════════════════════════════════════════════════════
→ POST /api/client/register
authorization: *:production.e4455...
unleash-appname: codeium-language-server
Body (561 bytes, JSON):
{"appName":"codeium-language-server","instanceId":"..."}
```
## Architecture
```mermaid
graph LR
A[Your App
OpenAI SDK] -->|HTTP| B[Proxy
:8741]
B -->|gRPC| C[Language
Server]
C -->|HTTPS| D[Google /
Anthropic]
E[MITM Proxy
:8742] -.->|intercept| D
C -.->|routed via| E
```
## CLI Flags
```
antigravity-proxy [OPTIONS]
Options:
--port API server port (default: 8741)
--no-mitm Disable MITM proxy
--mitm-port Override MITM proxy port (default: auto)
```
## License
Private. Do not distribute.