REST API

Gilial provides REST endpoints for compression operations.

Base URL

http://localhost:8000

(or your deployed endpoint)

Endpoints

GET /status

Get current index statistics.

Response:

{
  "index_name": "your-index",
  "environment": "us-west-1",
  "total_vector_count": 10000,
  "dimension": 768,
  "namespaces": {}
}

Example:

curl http://localhost:8000/status

GET /estimate

Estimate compression savings.

Query Parameters:

• strategy (optional): "balanced" or "aggressive" (default: "balanced")

Response:

{
  "original_vectors": 10000,
  "compressed_vectors": 9739,
  "original_size_mb": 29.30,
  "compressed_size_mb": 28.53,
  "savings_pct": 2.61,
  "compression_ratio": 0.974
}

Example:

curl "http://localhost:8000/estimate?strategy=balanced"

POST /compress

Compress vectors (dry-run or apply).

Request Body:

{
  "strategy": "balanced",
  "dry_run": true
}

Parameters:

• strategy (str): "balanced" or "aggressive" (default: "balanced")
• dry_run (bool): Preview only (default: true)

Response:

{
  "strategy": "balanced",
  "original_vectors": 10000,
  "compressed_vectors": 9739,
  "original_size_mb": 29.30,
  "compressed_size_mb": 28.53,
  "dry_run": false,
  "savings_pct": 2.61,
  "metadata": {
    "deleted_vectors": 261
  }
}

Example - Dry-Run:

curl -X POST http://localhost:8000/compress \
  -H "Content-Type: application/json" \
  -d '{"strategy": "balanced", "dry_run": true}'

Example - Apply:

curl -X POST http://localhost:8000/compress \
  -H "Content-Type: application/json" \
  -d '{"strategy": "balanced", "dry_run": false}'

Running the Server

python backend/main.py

Server starts at http://localhost:8000

Rate Limiting

No built-in rate limits. Recommended:

• Limit compression runs to once per hour
• Run during off-peak times
• Use dry-run before applying changes