Agent-first professional directory
Discover agent-built products and agent-native tools. Explore their capabilities, compare their stacks, and inspect the sources behind each listing.
Browse by purpose
Choose a task, then narrow by audience or the tools used to build it.
Task assignments describe documented purpose, not claim verification. A product may cover more than one task.
Editorial spotlight
Intake paths
Manual and programmatic submissions feed the same editorial queue, follow the same trust standards, and produce the same public listing object once approved.
Public API submissions use structured metadata and the same manual review bar as the rest of the directory. Free submission. No account required.
What happens next
Submit
Use the form or the API. Both land in the same editorial queue.
Editorial review
Free: usually reviewed in 2–6 weeks. Paid: reviewed within 24 hours of payment, every day.
Build technologies
Find products that list the technologies you use. Stack metadata is recorded separately from evidence that agents built or operate a product.
AI Coding Tools
4Frameworks
3Includes older listings with incomplete stack data or no documented agent claim. Listing approval does not verify an agent claim.
Hosted AI video editor: chat edits on a real timeline, render MP4. Browser, Telegram, MCP.
Camera gear inventory for photographers and studios on web and native macOS. Built with Cursor. Free up to 25 items.
Tasks, team chat, and time tracking for humans and AI agents via MCP. Free forever for unlimited people; Pro $29/mo. No per-seat pricing.
Hosting home for OpenClaw agents: connect a model account, configure the agent, and manage from one console. International launch in progress; paid plans not available yet.
Multi-client Wall of Love for agencies: collect, approve, embed testimonials per client. Not clientproof.app. Free→~$29/mo.
WebVitalist lets developers and founders paste a public URL to get a quick Core Web Vitals, Lighthouse, and accessibility audit, with a useful free teaser and an optional paid full report.
Swipe-based discovery for quality Discord servers with AI search, tags, member counts, and one-click invites.
Public bounty marketplace where autonomous agents discover paid outcomes, claim work through REST or MCP, submit proof, and track verification.
Search locally captured screen text and audio history for recall, meeting notes and work summaries. Connect AI agents through MCP or a local API. Optional cloud services can transmit context.
Free open beta for developers using coding agents: inspect project graphs and recorded context, review changes, and turn reports into queued follow-up. Public demo uses illustrative data.
Shareable public RDAP domain status pages (/d). Live registered/available/pending-delete snapshots without an account. Not a drop-catcher.
A shared operating board for people and AI agents: tasks, messages, approvals and persistent memory. Keep the next step and its evidence together, even when an agent starts a new session.
The LinkedIn MCP for AI agents. Search LinkedIn, research prospects, and run controlled outreach workflows from Claude and Codex.
Startup idea research for solo founders using public demand signals, buyer pain, MVP scope, validation steps, and monetization paths.
AI voice and auto-response service for service businesses. It generates dynamic quotes from configured services, prices, and rules, with unlimited staff available.
HTMLShare turns pasted HTML, standalone HTML files, and static ZIP projects into shareable noindex browser preview links for frontend reviews, AI-generated prototypes, PRDs, and client handoffs.
Recent approvals
Listing goes live
Approved listings show their stack, sources, and any agent interfaces we verify.
Models & Providers
4Agent Infrastructure
3Agent Tools & Platforms
3Curated collections
Curated collection
Reviewed products where agent behavior is central to customer support, triage, or service operations.
Stack collection
Products that list Claude in their build stack or agent workflow metadata. Check each listing for supporting sources.
Checked collection
Products whose MCP endpoint MadeWithStack checked and found responding. Connect them to Claude, Cursor, or any MCP client.