Store snapshots and page diffs
Connect
SPIRITT
TinyFishManagedNo API needed, pay per usage.
The live web.
Working for you.
Search the open web, news, and research papers with TinyFish, fetch pages as clean Markdown or structured JSON, and send browser agents through multi-step sites, while SPIRITT builds the systems that use the results.
What you can do with TinyFish.
Find what matters
Search the web, recent news, and research publications with domain, language, location, and date filters.
Read any page cleanly
Fetch up to ten URLs at once as Markdown, HTML, or structured JSON, with links, metadata, and selectors.
Agents that navigate
Run browser agents toward a goal in plain language, with step screenshots and HTML captured as evidence.
Works with your other apps.
Get started in three steps.
01
Open your workspace.
Open a SPIRITT workspace where your agent can search, read, and navigate the web and run your data workflows.
02
Describe the web data you need
Name the sites, topics, and goals, and where the results should land. No API needed, pay per usage.
03
SPIRITT builds and runs it
SPIRITT runs the TinyFish searches, fetches, and browser agents, structures the results, delivers them to your tools, and keeps it current.
Hand over a mission.
Choose a mission. Make it yours.
Put TinyFish to work.
Put the live web to work for you.
Questions
TinyFish, with SPIRITT.
01Do I need a TinyFish account or API key?+
No. No API needed, pay per usage. SPIRITT manages TinyFish access for you, so there is no TinyFish signup, plan, or credential to set up. Describe the sites, topics, or goals you care about and SPIRITT starts working.
02What can SPIRITT do with TinyFish?+
SPIRITT can search the open web, recent news, and research publications with domain, language, location, and date filters, fetch up to ten URLs at a time as Markdown, HTML, or structured JSON, and run asynchronous browser agents toward a natural-language goal, checking progress, retrieving step screenshots and HTML, and cancelling runs.
03When does SPIRITT use a browser agent instead of a fetch?+
Fetching is best for pages whose content loads directly. A browser agent is for multi-step work, such as filling a search form, paging through results, or clicking to reveal details. Agent runs take longer and cost more, so SPIRITT uses them only where a simple fetch falls short.
04Are there limits on what TinyFish can reach?+
TinyFish works with publicly reachable pages. Sites behind your private logins, paywalls, or aggressive bot protection may return partial results or fail. Search indexes reflect what TinyFish has crawled, so the newest items can lag. SPIRITT keeps source links and timestamps so every result is traceable.
05How does SPIRITT turn TinyFish results into working systems?+
SPIRITT schedules searches and fetches, structures the content, diffs changes over time, stores history in a database, and pushes summaries, records, and alerts into Notion, Slack, HubSpot, Salesforce, and more. It keeps these workflows running and repairs them when sites change.
06How is TinyFish usage billed?+
Usage is metered per call and drawn from your workspace credits. Managed access is not free or unlimited, and browser-agent runs cost more than searches or fetches. Failed provider calls are not billed.