the catalog keeps growing.

connect a plain key and your agent can work your support inbox: list and search threads, read a whole conversation, reply to the customer on whatever channel they used, mark threads done or snoozed, assign them, set priority, label them, and leave internal notes the customer never sees. customers can be looked up by email or your own external id, and upserted. anything outside that loop — companies, tenants, tiers and slas, tasks, snippets, the help center — goes through a raw graphql tool on the same key.

connect a dev.to key and your agent can write to your blog: publish or draft a post, edit one, take it back to draft, and read your posts, analytics, followers, followed tags and reading list. browsing articles, search and tags need no account at all. tags accept an array even though forem wants a comma string.

find where to launch a product: startup directories ranked by domain rating, each with its price and whether the link is dofollow, plus the weekly launch leaderboard and product search. scrolllaunch ships these as an mcp server with no rest api, so fetchbean calls the mcp tools for you and hands back plain json. no account needed.

scan any product url into a ready listing draft (name, tagline, descriptions, category, tags) with no account needed, and check the free launch-week calendar. connect your own key and your agent can draft, edit, submit for review, issue and verify the badge, and open featured checkout. free listings take a weekly slot and need a reciprocal badge for dofollow; featured is $11.99 one-time with neither.

check any site's ahrefs domain rating, the 0-100 log scale for backlink strength, plus the global top 100 by dr. works with no setup, or connect your own free ahrefs key for 5 credits a call instead of 20. the response carries the attribution ahrefs requires next to the number.

connect an ElevenLabs key to speak text in any voice, generate sound effects, and transcribe audio with Scribe (diarization and word timings included), plus your voices, models and remaining character quota. generated audio comes back as base64 and as a link fetchbean serves, so you can hand off a url instead of megabytes.

connect a Sarvam key for Indian languages: speak text with bulbul across 11 languages, transcribe calls with saaras, translate and transliterate across 20+ languages, detect a language and script, and chat with sarvam-m. text to speech returns Sarvam's own base64 response untouched, with archived links alongside it.

six more canny reads: list and retrieve ideas (with search and field filters), the insights behind an idea, the salesforce and hubspot deals linked to a request, the groups ideas are organized into, and companies with their spend. comments, votes and status changes now call canny's documented endpoints for those reads, which return richer filters and paginate by cursor instead of skip.

merging duplicate posts or ideas, and deleting a post, idea, comment, vote, category, company or user, are now named canny tools instead of raw passthrough calls. your key could always do this; now an agent can find it. none of it is recoverable, so each tool spells out what it takes with it.

connect a Loops API key to manage lifecycle campaigns, contacts, email components and themes, transactional messages, events, uploads, and complete workflow graphs. 63 named tools cover Loops' current OpenAPI surface, with a host-pinned raw request tool for future additions and Loops' own validation, limits, and errors passed back.

connect Resend with OAuth for send-only or full account access. send, schedule, inspect, and receive email; manage templates, contacts, broadcasts, domains, suppressions, segments, topics, webhooks, metrics, and logs. a raw request tool reaches the rest of Resend's API directly, with Resend's own validation and errors passed back.

connect one credential exported from Google's gws CLI, then use Gmail, Drive, Calendar, Docs, Sheets, Slides, Forms, Tasks, People, Chat, Classroom, Keep, Meet, Workspace Events, Apps Script, Admin Reports, and Model Armor through Google's REST APIs. fetchbean live-validates the refresh token and scopes, refreshes access directly with Google, reads live Discovery schemas, supports pagination and media, and can back the existing curated Google tools when their narrower connection is absent. no CLI runs on fetchbean servers.

connect discord, install the fetchbean bot into as many servers as you need, and choose a default for calls that omit guild_id. list and search messages, send and reply, manage channels and threads, react, pin, and join or leave supported servers. every bot action is checked against the calling organization's installed-server list before the shared bot token is used.

inspect and administer your tailnet with named tools for every stable write in tailscale's current api: devices, routes, posture, invites, logs, dns, keys, policy, users, contacts, webhooks, settings, services, oauth apps, and organization tailnets. connect an oauth client with only the scopes you want the agent to have; the raw request escape hatch remains for future api additions without allowing an off-host request.

run any model on replicate: generate images, video, audio and text, search the model catalog, and manage predictions and fine-tunes. `replicate.run` waits for the output and hands it back in one call; slow models come back still processing and you poll for them. connect your own replicate token and any model runs at the flat 5 credits, billed to your account. without one, model search and the flux image models still work on ours, metered per image you ask for.

your support inbox, on your own access token. read conversations, messages and contacts, reply or leave an internal note, resolve, snooze, assign and label, and pull first-response and resolution reports. works against chatwoot cloud or your own self-hosted install.

search your intercom conversations and contacts, reply as your teammate, close, snooze and assign, tag and note, file tickets, and search the help center before you answer. us, eu and au workspaces.

register a sitemap with bing (SubmitFeed) or unregister one (RemoveFeed). you could already read submitted sitemaps and submit individual urls; now you can register the sitemap itself instead of finishing by hand in the bing ui.

read the comment thread on an issue by uuid or identifier (ENG-123). you could already post comments; now you can read them too, without dropping to raw graphql.

three more linear reads: list a team's workflow states (the stateId to move an issue to), your workspace members (the assigneeId), and issue labels (the labelIds). the pieces you needed to triage an issue without dropping to raw graphql.

nine more canny tools: edit a post, tag and untag, move it to another board or category, list categories and status changes, look up or upsert a user. everything paginates the same way (limit + skip) apart from users, which canny only documents on its newer cursor-based endpoint. the connect form now checks your key is shaped right before you save it.

your feedback boards, on one key: read posts, votes and comments, search and filter by board, status or tag, file a request, move one along the roadmap, reply, and publish a changelog entry when it ships. voters are only emailed when you ask, and no tool can delete a customer's feedback.

your site's bing search performance: top queries and pages, keyword and related-keyword volume, crawl issues, backlinks, sitemaps, and submitting urls for indexing (connect with one click).

your agent can publish an md or interactive html page and get back a shareable link a human can open. create, update in place, roll back to a prior version, set unlisted or private, and manage them all from the dashboard.

brand search: resolve a company or brand name to its domain, over logo.dev's brand index.

jina now does web search too — ranked serp links (title, url, snippet), alongside the reader. shows up under web search.

your cal.com scheduling: list event types and availability, list bookings, and create or cancel bookings (bring your own api key).

classify any company into naics or sic industry codes (with confidence) by domain or company name. two new tools on context.dev.

brand intelligence for any company: logos, colors, fonts, description, socials, industry, and address by domain, email, name, or ticker.

scrape any url to clean, js-rendered markdown for script-heavy pages a plain reader misses, plus web search.

your render infrastructure: services, deploys, logs, metrics, workspaces, and deploy controls (bring your own render api key).

your google, as one-click connections: calendar (list/read/create/update/delete events), docs, sheets, slides, and forms (read and write), gmail (send email), analytics (ga4 reports), and search console (search analytics, sites, url inspection).

read and post to your bluesky account: home timeline, profile, notifications, post search, and create/delete posts, like, repost, follow (bring your own app password).

schedule and manage social posts across your connected channels: list channels, create/edit/schedule posts, and read analytics (bring your own buffer api key). heads up: buffer's api has a tight monthly request quota.

your clerk instance: users, organizations, sessions, and invitations (bring your own secret key).

track movies and tv on your own trakt account: watchlist, watched history, ratings, and search.

the movie database: search movies and tv, fetch details and images, trending, and recommendations.

your stripe account: customers, payments, subscriptions, invoices, and refunds (bring your own key).

your slack workspace: post and read messages, list channels and users (connect in one click).

your sentry org: list and inspect issues and events, and resolve issues (bring your own token).

your cloudflare zones, dns, and cache (bring your own scoped account token).

your supabase org: run sql on its projects, list projects, and read logs (one-click connect).

search, read, and update your notion pages and databases (connect with one click).

your gitlab projects, issues, merge requests, and pipelines (gitlab.com or self-managed).

your dodo payments account: payments, subscriptions, customers, and refunds (bring your own key).

your spaceship domains and dns: list domains, edit records, set nameservers and auto-renew.

your better stack uptime monitors, incidents, and logs (bring your own key).

github repos, issues, pull requests, and code search (public data, or your own token for more).

read and write your linear issues, projects, and comments (bring your own key).

your revenuecat subscriptions, customers, and product data (bring your own key).

your expo/eas builds, updates, submissions, and apps; cancel a build (bring your own token).

your meeting transcripts, summaries, and action items (bring your own fireflies key).

query your posthog events (hogql), feature flags, insights, and errors (bring your own key).

ios app store search, metadata, reviews, ratings, and ai review insights.

hosted utilities and data: domain and dns lookups, sitemaps, url metadata, quotes, facts, and more.

on-demand weather data by city or coordinates.

remote job listings across the web.

neural web search built for retrieval across the open web.

turn any url into clean, llm-ready markdown.

agent-focused web search with optional synthesized answers.

web search with synthesized answers.

multi-query web search and url extraction.

social data, including public instagram profiles.