
Haystack MCP server
Which MCP servers are Haystack'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 2026-11-03
- Registry nameio.haystackapp/jobsverified in the MCP registry
The registry name sits under haystackapp.io. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.
- Stateanswerschecked 2026-11-03
- Tools4
- Server calls itselfhaystack
- Server version1.0.0
- Protocol version2025-06-18
On 2026-11-03 the server answered initialize and tools/list without a sign-in. First answer in 400 ms.
- Whether its tools do the product's job has not been judged yet.
- Checked on 1 of the 7 days it takes to judge reliability.
Once connected
- Parameters are untyped or undescribed in 1 of 4 tools: check arguments before calling.
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 scoreafter a weeka score needs 7 daily checks: 1 so far
- Answeredmeasuring1 of the 7 daily checks it takes to print a share
- Median first answer1530 ms
- Space in an agent's context≈ 724 tokensthe definitions of all its tools (4), which an agent carries on every turn
- Tool definitionsA · 96 of 100
- Typed results0 of 4 toolsdeclared outputSchema: results come typed, read once
- Parameters without a description: in 1 of 4
What an agent can do with it
- not sure · Match a resume to live job postings
- not sure · Have an agent apply with a tailored resume
- cannot · Post a role and screen incoming applicants
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_jobsSearch live jobs on Haystack, newest first. Filter by job title words, country, city, remote, and by what the advert says: no experience needed, no degree needed, visa sponsorship offered. Returns up to 20 jobs, each with its link, and an…
get_jobOne job's details by its Haystack id: title, employer, location, pay, work mode, skills, what the advert says (no experience, no degree, visa sponsorship), the description, and facts about applying to that employer from Haystack's data…
employer_factsWhat Haystack knows about applying to an employer: live roles, the share of applications that got a reply, the median time to the first reply, how many fields its form has and the questions it asks. Pass no company to get the employers…
application_system_factsFacts about an application system (ATS) from applications Haystack's AI agent sent: how many fields its forms have, the questions they ask most, and how fast employers on it reply. Without a system, the overall figures.
How to connect
Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:
{
"mcpServers": {
"jobs": {
"type": "http",
"url": "https://haystackapp.io/mcp"
}
}
}Add it
Claude Code, in a terminal
claude mcp add --transport http haystackapp https://haystackapp.io/mcpclaude.ai, ChatGPT and other clients: add https://haystackapp.io/mcp as a custom connector in the settings.
- io.haystackapp/auto-applyanswers, requires sign-inverified in the MCP registry
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.