TinyFish
TinyFish provides web automation, live search, content extraction, research, browser sessions, reusable browser profiles, and run management through native REST APIs.
Verdict
Common use cases
- Scrape competitor pricing from JS-rendered pages
- Monitor site changes behind login walls
- Test checkout flows across browser profiles
- Extract job postings from dynamic career sites
- Audit SEO metadata on client-side apps
Integration
- Vendor
- TinyFish
- Category
- other
- Auth
- API_KEY
- Tools
- 23
- Composio slug
tinyfish
Tools
- Cancel Profile Setup Session
Cancel a Browser Context Profile setup session and permanently discard its unsaved cookies and browser storage.
- Cancel Runs
Idempotently cancel one or up to 100 asynchronous or SSE runs. Active cancellation is irreversible; synchronous runs cannot be cancelled, and already-terminal runs are returned with their actual status.
- Create Browser Session
Create an isolated TinyFish remote browser session and return its CDP connection URL. This action starts billable browser infrastructure: TinyFish meters Browser usage until the session ends or becomes inactive. You must call Terminate Brow
- Create Profile
Create a reusable Browser Context Profile for cookies and browser storage; this changes persistent account state.
- Delete Profiledestructive
Permanently delete a Browser Context Profile and all browser state saved in it. This action cannot be undone.
- Delete Profile Domaindestructive
Permanently remove saved cookies and browser storage for one domain from a Browser Context Profile. This action cannot be undone.
- Fetch URLs
Render and extract content from 1-10 URLs in one provider-native batch. Always inspect both results and errors because HTTP 200 can contain successful pages and per-URL failures together.
- Get Profile
Get one Browser Context Profile and its saved-domain metadata.
- Get Runs
Get summary records for 1-100 TinyFish automation run IDs in one request. The response separately reports IDs that were not found or are not owned.
- Get Wallet
Get the TinyFish wallet balance, auto-reload and pending-top-up state, and current Agent and Browser rates. Monetary values are returned as exact decimal strings. Legacy-credit accounts report FEATURE_NOT_AVAILABLE rather than an authentica
- List Profiles
List Browser Context Profiles available to the connected TinyFish account. Use the returned IDs when attaching or managing reusable browser state.
- List Research Runs
List saved TinyFish research runs with optional status, query, date, and sort filters, returning one cursor-controlled page.
- List Runs
List automation runs with optional status, goal, date, and sort filters, returning one cursor-controlled page.
- List Usage
List one page of Search, Fetch, or Browser usage history. Search and Fetch usage reads are free; Browser usage requires the paid Browser service.
- List Vault Connections
List display-safe metadata for password-manager vaults connected to TinyFish. Credentials and secret values are excluded.
- List Vault Items
List display-safe metadata for items synchronized from password-manager vaults connected to TinyFish. Secret field values are never returned.
- Save Profile Setup Session
Persist cookies and browser storage from a completed setup session into a Browser Context Profile, then clean up the temporary session. Persisted state and successful session cleanup cannot be undone. The response retains domain-level skips
- Search Web
Search live web pages, news, or research papers with domain and time filters, optionally fetching result pages inline, and return exactly one caller-selected zero-based result page.
- Start Profile Setup Session
Start a billable temporary remote browser session for establishing login state in a Browser Context Profile. The returned browser state is not saved automatically: after setup, always call Save Profile Setup Session to persist it or Cancel
- Sync Vault Items
Refresh display-safe item metadata from all password-manager vaults already connected to TinyFish. Synchronization can add, update, or remove mirrored items. Returns synchronization counts and allowlisted item metadata without credential or
- Terminate Browser Session
Idempotently terminate a remote TinyFish browser session to stop further browser usage charges. Already-ended sessions are also successful.
- Update Profile
Rename a Browser Context Profile or change its default and proxy-country settings. Omitted fields remain unchanged.
- Upload Profile State
Upload cookies and local or session storage directly into a Browser Context Profile. This changes persistent login state and may transmit authentication secrets; call it only with the credential owner's explicit authorization. TinyFish prov
Setup
Setup guide
- 11. Log into your TinyFish dashboard and generate a new API key from the credentials or settings page. 2. In Switchy, open your workspace settings and navigate to Integrations, then click Add Integration and select TinyFish. 3. Paste your API key into the authentication field and click Connect. 4. Switchy will validate the key and confirm the connection — you'll see TinyFish appear in your list of active integrations. 5. Open any Project, type '@TinyFish fetch the homepage of example.com' in a message, and send it. 6. The AI will invoke TinyFish's fetch tool and return the rendered HTML or extracted content, confirming the integration works. 7. To test profile management, try '@TinyFish create a profile named demo-profile' and verify it returns a profile ID you can reference in later prompts.
What teammates see: by default, memories from TinyFish are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Batch Fetch Competitor Pages
@TinyFish fetch the following URLs and extract their main content: https://competitor-a.com/pricing, https://competitor-b.com/pricing, https://competitor-c.com/pricing. Return the pricing tiers and feature lists from each.
Create Logged-In Profile
@TinyFish create a new browser profile named 'linkedin-scraper' and start a setup session. Give me the CDP URL so I can log in manually and save the session.
Monitor Page for Changes
@TinyFish fetch https://status.example.com and extract the current incident list. Compare it to the snapshot from yesterday and highlight any new entries.
Extract Job Listings
@TinyFish fetch https://careers.example.com/openings and extract all job titles, locations, and application links. Return the results as a table.
Cancel Long-Running Fetches
@TinyFish cancel all active runs that have been running for more than 5 minutes. List the run IDs and their original URLs.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@TinyFish I need to scrape product listings from three competitor sites daily. Set up profiles for each domain so we maintain consistent sessions across runs.
Created three Browser Context Profiles: 1. Profile 'competitor-alpha' (prof_x7n2k) — ready for alpha-products.com 2. Profile 'competitor-beta' (prof_m9p4j) — ready for beta-market.io 3. Profile 'competitor-gamma' (prof_r5t8w) — ready for gamma-shop.net Each profile will persist cookies and browser storage across sessions. When you fetch URLs or create browser sessions, reference these profile IDs to maintain login state or bypass bot-detection fingerprints. Profiles remain active until explicitly deleted.
This showcases TinyFish's profile management for stateful scraping. The Create Profile tool lets you save browser context (cookies, localStorage) and reuse it across runs—essential for authenticated scraping or avoiding repeated CAPTCHAs. Trade-off: Profiles are persistent account state, so remember to Delete Profile when a workflow ends to avoid clutter and potential stale-auth issues.
Use-case deep-dives
When TinyFish beats headless Puppeteer for weekly price checks
A 3-person DTC brand runs weekly competitor pricing audits across 40 SKUs on sites that block standard scrapers. TinyFish's remote browser sessions handle the JavaScript rendering and anti-bot evasion without maintaining your own Puppeteer cluster. The batch Fetch URLs tool pulls 10 pages at once, and reusable Browser Context Profiles store login cookies so you're not re-authenticating every run. This works until you hit 200+ URLs per week—at that scale, the per-session metering cost outweighs spinning up your own headless infrastructure. If your scrape is weekly and under 100 URLs, TinyFish keeps the ops burden near zero. Set up the profile once, schedule the fetch, and parse the results in Switchy.
Why TinyFish fits end-to-end test orchestration in Switchy
A 6-person product team runs post-deploy smoke tests on their multi-tenant SaaS: log in as three different user roles, verify dashboard load, check feature flags. TinyFish's Create Browser Session and Profile tools let you script these checks in Switchy without a dedicated CI runner or Playwright config. Each profile holds one test user's cookies, and the CDP connection URL feeds real browser events back to the AI for pass/fail logic. The trade-off: if your test suite grows past 20 scenarios or needs pixel-perfect screenshot diffs, you'll want a proper testing framework. For 5-10 critical-path checks that a non-engineer can read and edit in Switchy, TinyFish keeps QA in the same workspace as your product docs and runbooks.
When TinyFish solves the paywalled-data problem for sales ops
A 4-person sales team qualifies inbound leads by checking their company's tech stack on sites like BuiltWith or Crunchbase, which require logins and rate-limit APIs. TinyFish's Profile system stores the login session, and Fetch URLs pulls the enrichment data without hitting API quotas. The AI parses the rendered HTML and writes findings into your CRM. This breaks down if you're enriching 500+ leads a day—browser sessions are slower and pricier than native APIs. For 20-50 leads a week where the data lives behind a login or paywall, TinyFish turns a manual copy-paste task into a one-click Switchy prompt. Create the profile, save the login, and let the AI handle the rest.
Frequently asked
What does the TinyFish MCP do in Switchy?
TinyFish gives your AI agents remote browser sessions and persistent browser profiles. Agents can render JavaScript-heavy pages, manage cookies across sessions, and extract content from sites that block headless scrapers. It's built for workflows where you need real browser behavior — login state, session cookies, or client-side rendering — not just static HTML fetching.
Do I need special credentials to connect TinyFish?
You need a TinyFish API key. Generate one in your TinyFish dashboard, paste it into Switchy's connection flow, and you're done. No OAuth dance, no admin approval. The key controls billing and rate limits on TinyFish's side, so treat it like a production secret and rotate it if anyone leaves the team.
Can TinyFish MCP interact with pages after they load?
No. TinyFish renders pages and extracts content, but it doesn't click buttons or fill forms through the MCP. If you need multi-step browser automation — login flows, form submissions, pagination — you'll want Playwright or Puppeteer running in your own infrastructure. TinyFish is optimized for fetch-and-extract, not interactive scripting.
Why use TinyFish instead of calling the API directly?
The MCP wraps TinyFish's batch-fetch and profile-management endpoints so your AI agents can invoke them without you writing glue code. Agents see tool descriptions, handle errors, and chain browser sessions with other MCP integrations. If you're already piping API responses into prompts manually, the MCP just removes that boilerplate.
Who on the team should connect the TinyFish MCP?
Whoever owns your TinyFish account and understands your scraping budget. Browser sessions are metered and profiles persist across runs, so the person connecting it should know which domains you're targeting and how much monthly usage you've allocated. Once connected, any Switchy workspace member can invoke the tools in their agent workflows.