How TeCoxBeta works

Search is only the beginning.

TeCoxBeta turns web discovery into a clear research flow: understand the question, gather evidence, read what matters, and return a result that people can use.

01 · UNDERSTAND

Plan the question

Translate a broad question into relevant angles for discovery.

02 · DISCOVER

Search in parallel

Collect results across the enabled source types and providers.

03 · EVALUATE

Rank and read

Combine signals, choose promising sources, and extract meaningful page content.

04 · EXPLAIN

Return an answer

Deliver sources, citations, events, telemetry, and an answer length chosen by the user.

Evidence stays visible

Research results should be inspectable.

A final answer is useful when people can see the source set behind it. TeCoxBeta returns source metadata, rankings, citations, and useful pipeline diagnostics beside the written result.

  • Use /sources for retrieval-only workflows.
  • Use /search for JSON or streamed search.
  • Use /health and /providers for operations checks.
response.json200 OK
{
  "answer": "…[1] [2]",
  "max_output_tokens": 1200,
  "sources": [ ... ],
  "provider_stats": { ... },
  "elapsed_ms": 4821.7
}

Instant through Ultra

Depth controls let the caller choose how much retrieval, reading, and synthesis work belongs to the request.

Answer length is separate

Use max_output_tokens when the research should remain deep but the response should remain focused.

Ready to use the workflow?

Start with the workspace, then move to the API when you are ready to build.