FindAll supports discovering companies and people based on natural-language objectives and custom match conditions. You can define criteria like industry, employee count, location, funding stage, or any attribute expressible in natural language. The API returns structured profiles with matched and unmatched candidates up to 1,000 results per run.

Parallel's Search API runs multiple queries simultaneously and returns semantically ranked documents rather than keyword-matched links. Your agent can batch several related queries into one call, getting top-matching results across all of them. This means richer, more relevant answers synthesized from multiple data sources in a single response.

No. For complex FindAll runs, the agent submits the job and uses Parallel's status endpoint to poll progress. It can inform the customer that research is underway, provide intermediate updates, and deliver final results when ready. Task groups also stream events so the agent can relay completions incrementally.

Tars requires your Parallel API key, which you can generate from your Parallel dashboard. Enter it once in the Tars Tools section. The key is stored securely and never exposed to end users. You can rotate or revoke it anytime from the Parallel console.

Yes. Parallel's FindAll API accepts a generator parameter that sets the quality and cost tier for each run. You can configure your agent to use different tiers based on the request type, balancing depth of research against credit consumption for routine versus high-stakes queries.

Tars does not maintain a separate database of Parallel results. Research data is fetched live during conversations and used to formulate responses. Conversation logs may contain the summarized output, but raw Parallel payloads are not persisted beyond the session unless you configure an export.

Using Parallel through Tars means your team interacts with research capabilities conversationally instead of writing API calls. The AI agent translates natural questions into properly formatted FindAll objectives, search queries, and task groups. Non-technical team members get the same structured intelligence without touching code or dashboards.

Absolutely. Your agent can run a Parallel FindAll to discover target companies, then cross-reference results with your CRM data from HubSpot or Salesforce, enrich contacts through other integrations, and present a unified view. Tars orchestrates multiple tool calls within a single conversation flow.