Kado
kado.so- Category
- Developer Tools
- Rank
- No. 2330Tools index
Previous survey · No. 2205 ·
- Type
- TOOL
- Use case
- Data, Retrieval & Knowledge
- Date
About
Kado is a search engine built for AI agents rather than people, indexing roughly 3,000 MCP servers and A2A agent endpoints by using LLMs to model what outcome each one actually delivers, so a personal agent can find the right "agent app" to delegate a task to instead of relying on regular web search. The index was pruned from about 45,000 entries down to ones an agent can actually invoke, and retrieval is based on a modified HyDE approach the team is working to speed up from its current multi-second latency.
At a glance
Kado gives agents a search endpoint that returns findings and linked sources. It is worth evaluating if you want a quick starting point for public-web research and can check the original pages yourself.
Good fit
Developers testing search integration for an agent that needs citations.
Consider something else when
Use a dedicated authoritative index when coverage or reproducibility is essential.
Start here
Try a narrow public question whose answer you know. Open two cited pages and check whether the dates and claims support the returned answer.
Cost and requirements
The site advertises free agent search and links to API and MCP documentation. Its earlier Kado Direct product is marked sunset.
Limits to know
Returned summaries depend on retrieved pages and synthesis. Check current access terms and avoid private queries until the data path is clear.
Sources checked 2026-11-05. Source-based assessment.
Why it made the leaderboard
Kado indexes thousands of MCP and A2A agent apps so a personal agent can find and delegate to the right specialized agent instead of relying on generic web search.
Tags
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.