RecruitingConnector
Ashby
Search and read Ashby candidates, jobs and applications, and add notes to a candidate, from Claude or ChatGPT through Tulina's MCP server and an Ashby API key.
Official site: www.ashbyhq.com
What it can do
2 read/1 write/1 send
- 01Send
Search candidates by email or name, and read a candidate
- 02Read
List and read jobs, filtered by Open, Closed, Draft or Archived
- 03Read
List and read applications, filtered by job
- 04Write
Add a note to a candidate
The connector reads the hiring funnel and writes one thing back: a note on a candidate. Candidates, jobs and applications are read-only, so it suits reporting, preparation and record-keeping around the pipeline, and it does not move a candidate between stages or create one. It authenticates with an Ashby API key created by an admin.
Does Ashby have an MCP?
Yes. Ashby hosts its own server at https://mcp.ashbyhq.com/mcp/v1. Each user signs in with their own Ashby account through OAuth, and an Org Admin first switches the server on under Admin, Organization Setup, Opt-In Features. Ashby's documentation lists ChatGPT, Claude, Cursor, Glean and Gemini CLI among the clients, says it is available on all plans except Analytics-only organisations, and says access inherits each user's permissions. As of October 2026 the page is titled beta and warns that tool inputs and outputs may change without notice.
How do you connect Ashby to Claude and ChatGPT?
Through Tulina's MCP server, in three steps:
- In Ashby, open Admin, Integrations, API and create a key.
- Paste the key into Tulina's Ashby connector, for yourself or for the organisation. It is stored in the vault and never shown to the model.
- Add
https://mcp.tulina.ai/mcpto Claude or ChatGPT and sign in with your Tulina account. The MCP page has the same link.
Several keys can be connected as named accounts, and a request names the one it is for.
Tulina vs Ashby's official MCP server
Both reach your Ashby data. They differ in who signs in, and in how much each can change.
| Ashby's MCP server | Tulina | |
|---|---|---|
| Endpoint | mcp.ashbyhq.com/mcp/v1, hosted by Ashby | mcp.tulina.ai/mcp, one link for every connected tool |
| Sign-in | OAuth per user, switched on by an Org Admin | An API key created in Ashby, held in Tulina's vault |
| Whose permissions | Each user's own, so it cannot return what that user cannot see | The key's |
| Reading | Candidate profiles with resume, feedback and notes, filtered searches with counts, interview plans, upcoming interviews, pipeline counts per stage | Search by email or name; list and read candidates, jobs and applications |
| Writing | Create a candidate, add a note, move an application to another stage, consider a candidate for a job | Add a note to a candidate |
| In the same conversation | Ashby | Ashby plus your inbox, calendar, Slack and the other connected tools |
If the model should move candidates between stages, create them or prepare an interviewer, Ashby's own server is the better choice. Tulina fits when the pipeline is read as one input to something else, such as a Slack digest or a note filed where the team already works.
What you can ask it
"Find the candidate whose email is ada@example.com and show her applications." The agent searches candidates by email, reads the candidate, then lists applications for the job she applied to.
"Which jobs are open, and how many applications does each hold?" The agent lists jobs with the Open status, then lists applications per job and counts them. Both calls page with a cursor, so a long list is read in pieces.
"Add a note on Ada: strong systems design, follow up on salary expectations." The agent writes the note to the candidate's profile. No tool is dedicated to reading notes, so the first note on a new workflow is worth checking in Ashby.
A hiring read pairs naturally with Greenhouse or Lever for teams that run more than one tracker, and with Slack for a weekly pipeline post.
What the Ashby connector can't do
- It does not create candidates, move applications between stages or change a job. The only write is a note.
- It does not schedule interviews or read scorecards and interview feedback.
- It cannot edit or delete a note once written. The tool only adds.
- Candidate search matches by email and by name. It has no filter on stage, source or tag.
- Candidate and application lists are paged with a cursor, so a large pipeline is read in pieces.
- The key sees what its Ashby permissions allow, and every person using the connector shares them.
- It does not run under each user's own Ashby login. Access is the key's, which is a reason to create it with an admin's care.
Why through Tulina
The connection was never the hard part.
Most systems now publish their own way for a model to reach them, and that is good: it is what Tulina connects to. The hard part is deciding who may use it, what it is allowed to return, and what the model knows before it asks.
The keys stay in one place
YouTeamOrganisation
Connect Ashby once. The credential is encrypted in a vault, never held in clear, and never pasted into a chat or a workflow file. A person uses their own key where they have one, their team's where they do not, the organisation's after that, so nobody is blocked and nobody holds more access than they need. Revoke it once and everything running on it stops.
What leaves is filtered before it leaves
fields droppedmasked or hashedpseudonymised
Roles decide who can reach Ashby at all. Beyond that, a policy on the connector itself decides what comes back. The data is minimised at the source rather than after it has already reached a model.
Every call is on the record
whatwhowhen
Runs are logged for the organisation. That is what makes an automated action reviewable after the fact, and it is what a security review asks for first.
The run starts knowing your company
namingkey accountsdecisions
A direct connection gives a model access. It does not tell it how your company names things, which accounts matter, or what was decided last quarter. Tulina carries that across every tool, so the same request against Ashby returns something usable instead of something generic.
Processes that run on Ashby
None published yet.