Context API

High-quality context for your AI agents. One API, global caching, instant results.

Quick Start

Get your API key from the Dashboard and start making requests in seconds.

Authorization: Bearer sk-prismer-...POST /api/context/load

Global Cache

Shared across all agents worldwide. Pay once, use forever.

Instant Response

Cache hits return in ~50ms. No processing delays.

Smart Detection

Auto-detects URL vs query. One endpoint for everything.

API Endpoints

POST

/api/context/load

Smart context loader - handles single URL, batch URLs, or search queries automatically.

Single URL:{ "input": "https://example.com/article" }

Batch URLs:{ "input": ["https://example.com/1", "https://example.com/2"] }

Query Search:{ "input": "latest AI developments 2024" }

Parameters

Field	Type	Required	Description
input	string \| string[]	required	URL, URL array, or search query
inputType	"auto" \| "url" \| "urls" \| "query"	optional	Force input type (default: auto)
processUncached	boolean	optional	Process uncached URLs in batch mode (default: false)
search.topK	number	optional	Search results limit (default: 15, max: 30)
processing.maxConcurrent	number	optional	Max parallel compressions (default: 3)
processing.strategy	string	optional	Compression strategy (default: auto)
return.topK	number	optional	Return results limit (default: 5, max: 20)
return.format	"hqcc" \| "raw" \| "both"	optional	Response format (default: hqcc)
ranking.preset	"cache_first" \| "relevance_first" \| "balanced"	optional	Ranking strategy (default: cache_first)

Example

Request

{
  "input": "https://www.figure.ai/news/helix"
}

Response

{
  "success": true,
  "mode": "single_url",
  "result": {
    "url": "https://www.figure.ai/news/helix",
    "title": "Helix: Vision-Language Action Model",
    "hqcc": "# Helix: Vision-Language Action Model\n\n## Overview\n...",
    "cached": true,
    "cachedAt": "2026-01-02T10:30:00Z"
  },
  "processingTime": 45
}

Try it out

Request Body

Response

Click Execute to test the API

POST

/api/context/save

Store processed content in the global cache. Supports single or batch saves.

Single:{ "url": "...", "hqcc": "..." }

Batch:{ "items": [{ "url": "...", "hqcc": "..." }, ...] }

Error Codes

Code	Description
200	Success
400	Invalid request (missing or malformed parameters)
401	Invalid or missing API key
404	Context not found (cache miss)
429	Rate limit exceeded
500	Server error