Intelligence frontier
View moreMost popular LLMs and Tools
Sample market activityHighest spend
USDMost used
Runs- Routed models
- 9 9 with cost + quality evidence
- Callable tools
- 30 28 live now
- Core verified
- 31 Curated catalog entries
- Settled calls
- 0 Receipt-backed outcomes
Market map
Browse all 53 options Explore the catalog
- Tools
- 30
- LLMs
- 15
- Plugins
- 7
- Apps
- 1
Curated supply
Browse all tools Capabilities worth putting to work
01 · First-partyReady to hire
Fetch Page
by core · Tool
Fetch a public web page: title + readable text extract. Metered by bytes fetched.
- Ceiling
- ≤ $0.10/call
- Trust
- 100
- Calls
- New
Pinned and callable
02 · First-partyReady to hire
Sign Document
by core · Tool
Cryptographically sign a document (ES256). Returns a detached signature bundle as a file.
- Ceiling
- $0.05/call
- Trust
- 100
- Calls
- New
Pinned and callable
03 · First-partyReady to hire
Generate Image
by core · Tool
Create an original PNG with GPT Image 2 and save it to the agent's Files.
- Ceiling
- ≤ $0.30/call
- Trust
- 100
- Calls
- New
Pinned and callable
Receipt-backed signal
Open tool leaderboard New and notable
01Deep SearchResearch a question deeply and return a grounded answer with citations.≤ $1.00/call10002Search WebSemantic web search with category and domain controls.≤ $0.25/call10003Extract Web ContentExtract readable content and focused highlights from public URLs.≤ $0.25/call10004Find Similar PagesFind pages semantically similar to a public source URL.≤ $0.25/call10005Browse WebRender a public page in an ephemeral browser and extract information.≤ $0.50/call100
Evidence chain
Every recommendation has a paper trail.
Catalog identity, runtime behavior, and payment outcomes stay separate—and verifiable.
Definition-pinned
Schema drift refuses the call before money moves.
Trust-scored
Reachability and settled outcomes—not supplier claims.
Kernel-settled
Observed cost is capped; failed calls void at $0.00.