
AgentMash MCP server
Which MCP servers are AgentMash's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.
read 2026-10-28
- Registry nameai.agentmash/agentmashverified in the MCP registry
The registry name sits under agentmash.ai. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.
- Stateanswerschecked 2026-10-28
- Tools6
- Server calls itselfagentmash
- Server version0.1.0
- Protocol version2025-06-18
On 2026-10-28 the server answered initialize and tools/list without a sign-in. First answer in 130 ms.
- Its tools do get answers from saves with a link to each, but not share the best saved posts with teammates.
Once connected
- For "get answers from saves with a link to each", the likely first call is search_saves.
Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.
- INITE MCP score77 of 100answers 100 · does the job 50 · definitions 97 · context 100 · use 0 · connects 100
- What an agent can do with it1 of 2 use cases
- Answered100% of days7 of 7 daily checks
- Median first answer185 ms
- Space in an agent's context≈ 1,457 tokensthe definitions of all its tools (6), which an agent carries on every turn
- Tool definitionsA · 97 of 100
- Typed results0 of 6 toolsdeclared outputSchema: results come typed, read once
- Parameters without a description: in 1 of 6
What an agent can do with it
- not sure · search saved posts from an AI agent via MCP
- can do · get answers from saves with a link to each
- cannot · share the best saved posts with teammates
Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?
Worth a look
None of the five warning signs in these definitions.
A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →
search_savesSearch posts already imported into the authenticated user's AgentMash library. Use for requests to find their saved content by topic, author, platform (including newsletters forwarded to their Pro address) or collection. Matches keywords…
get_saveRead one saved post in the authenticated user's AgentMash library using an id returned by a saves tool. Returns stored full text, original URL, and available transcript, media, links, stats and collections. Missing content is not fetched…
recent_savesList posts imported into the authenticated user's AgentMash library within the last N days, newest imports first. Use for a recent library recap. Dates reflect import time, which can differ from when the user originally saved a post…
list_sourcesRead the authenticated user's AgentMash source status, saved-post counts, last sync times and reconnection status. Includes the newsletters source when forwarding is in use. Use to check which imported sources are available. Does not…
list_collectionsList the authenticated user's AgentMash collections with ids, names, sources and saved-post counts. Use to find a collection to browse. Includes imported folders and lists when available. Empty collections are omitted unless include_empty…
get_collection_savesList saved posts from one collection in the authenticated user's AgentMash library, newest imports first, using a collection_id from list_collections. Supports limit and offset pagination. An empty collection returns no saves; an unknown…
How to connect
Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:
{
"mcpServers": {
"agentmash": {
"type": "http",
"url": "https://agentmash.ai/mcp"
}
}
}Add it
Claude Code, in a terminal
claude mcp add --transport http agentmash https://agentmash.ai/mcpclaude.ai, ChatGPT and other clients: add https://agentmash.ai/mcp as a custom connector in the settings.
Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.
Before we count this visit
Analytics and ad measurement tell us which pages and which of our ads do their job: Google Analytics and Google Ads, the Meta pixel, Yandex Metrica, and one first-party id of our own. Nothing is stored until you answer, and declining changes nothing about the site.