
by Parallel
The Parallel Task API runs deep research and data enrichment in the background. It returns structured results for each run and offers processors from fast to very deep.
APIUsage-based
Step 1: Create a Parallel API key and set it as PARALLEL_API_KEY.
Step 2: Install the Python or TypeScript SDK, parallel-web.
Step 3: Choose a processor that matches how deep and fast the work needs to be.
Step 4: Submit a Task run and collect the structured result when it completes.
Step 5: Use the API chooser if you are unsure whether Task, Search or Responses fits.
From about ten seconds on the fastest processor to up to two hours on the deepest, because runs are asynchronous.
No. Pricing is per run, not per output field.
Task pricing is listed per 1,000 successful runs, with one charge per run regardless of output fields.
Search is synchronous and returns excerpts in 200 ms to 3 s for latency-sensitive queries, while Task is for deeper background work.
Perplexity Search APIThe Perplexity Search API returns real-time, ranked web results as JSON, with page text extracted into each result. It is for developers who want raw search results to feed their own models, agents or pipelines.
Trustworthy Language ModelTrustworthy Language Model is an AI-native product for API for producing language-model responses with confidence and trustworthiness signals for higher-stake.
Congress.gov APIThe Congress.gov API gives public programmatic access to congressional data, including bills, amendments, members, committees and House votes.