What's shipping

Launch dozens of agents in parallel and coordinate them with a first-class orchestration API. Large-scale scrapes that used to take hours now finish in minutes.

Watch your agents run in real time from the dashboard. Click in to take manual control at any point, then hand execution back to the agent.

Pre-warmed browser pools cut average session initialisation from 3.1 s to under 1.2 s across all regions.

Resolved edge-cases on certain CDN-fronted domains. Solver success rate is now 99.2%.

Push agent output directly into HubSpot, Salesforce, Slack, Notion, Airtable, and 13 more. No glue code needed.

One-click deploy for lead gen, competitor monitoring, SEO audits, social listening, and more.

Full TypeScript generics, improved typed errors, and a new streaming API so you can react to agent output as it happens.

Cookies are now reliably persisted across agent restarts in long-running authenticated workflows.

Track how your brand appears in AI-generated search results across Google, Perplexity, and ChatGPT search — the new frontier of organic visibility.

Use Airtap as a tool inside Claude Code or the OpenAI function calling interface. Describe a scrape in your prompt; Airtap handles the browser.

Residential proxy coverage expanded, with city-level geo-targeting now available for 120+ metro areas.

Fixed a memory leak in the session list view that caused slowdowns for accounts with 500+ concurrent sessions.

Airtap has completed its SOC 2 Type II audit. Security reports available to Enterprise customers on request.

Enable HIPAA-compliant data handling for healthcare workflows. Full audit trails, BAA available on request.

See compute minutes, API calls, and cost per team member live in the dashboard. Set per-project spending caps.

Webhooks now retry up to 5× with exponential back-off on 5xx errors. Failure logs retained for 30 days.

Drag-and-drop canvas for chaining AI actions, conditionals, loops, and transforms. No code required.

Agents now find buttons, inputs, and links using semantic understanding rather than brittle CSS selectors — surviving redesigns automatically.

Extract tables and nested data as clean JSON with automatic schema detection. Supports infinite scroll and multi-page pagination.