Search
Response Formats
Choose the right output format — full JSON, compact markdown, or YAML.
Response Formats
QANATIX supports three response formats. Choose based on your consumer.
json (default)
Complete JSON response. Best for applications.
{"query": "M8 bolt", "format": "json"}{
"results": [
{
"entity_id": "550e8400-...",
"name": "Stainless Steel Bolt M8x40 A2",
"score": 0.87,
"vertical": "manufacturing",
"entity_type": "fastener",
"vertical_data": {
"part_number": "SS-M8-40-A2",
"material": "Stainless Steel A2",
"price_eur": 0.12,
"in_stock": true
}
}
],
"metadata": { "..." },
"pagination": { "..." }
}~800 tokens per result.
compact
Markdown table. Best for LLM context windows. Returned as plain text.
{"query": "M8 bolt", "format": "compact"}| # | Name | Score | Key Data |
|---|------|-------|----------|
| 1 | Stainless Steel Bolt M8x40 A2 | 0.87 | part_number: SS-M8-40-A2, material: Stainless Steel A2, price_eur: 0.12 |~120 tokens per result. Use this for AI agents to minimize token costs. The MCP server uses this format automatically.
yaml
Abbreviated YAML. Best for readability. Returned as plain text.
{"query": "M8 bolt", "format": "yaml"}- name: Stainless Steel Bolt M8x40 A2
score: 0.87
part_number: SS-M8-40-A2
material: Stainless Steel A2
price_eur: 0.12
in_stock: true~200 tokens per result.
Comparison
| Format | Tokens/result | Response type | Best for |
|---|---|---|---|
json | ~800 | application/json | Applications, structured processing |
compact | ~120 | text/plain | LLM context windows, MCP, cost optimization |
yaml | ~200 | text/plain | Human readability, debugging |