BrightHire
by BrightHire
Query interview transcripts, scorecards, AI notes and coaching reports from your AI assistant. 11 read-only tools, OAuth sign-in, and a single read-suffixed scope that corroborates the read-only label.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Connect BrightHire via MCP
https://app.brighthire.ai/mcp/v1Works in any MCP-compatible client. In Agentman Studio it is one click — no config file to edit.
BrightHire Tools & Capabilities (11)
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
Limits
- One scope, and it is
.all. There is no way to grant narrower access at the OAuth layer — no split between transcripts, scorecards and coaching. Authorisation is read-everything-you-can-see. get_transcriptsreturns verbatim text. BrightHire describes the payload as large and advises using it only when exact wording matters. Twenty complete interviews can enter a model's context in one call.- No safety annotations were readable. The endpoint is gated, so we could not read
readOnlyHintordestructiveHintfor any tool. The read-only conclusion rests on the vendor's documentation and the scope, not on annotations. - BrightHire Screen data is excluded. BrightHire states the connector does not currently provide access to Screen product data, so AI-conducted screening interviews are out of reach.
- Coaching is self-only, as documented.
get_coaching_reporttakes no arguments and infers the caller. BrightHire documents no way to request another person's report through MCP. - An admin must enable the server. BrightHire documents an explicit enable step, so a user cannot connect unilaterally. That is a control, but it also means setup can stall outside your team.
- The one-click Claude connector is marked coming soon. BrightHire's own documentation flags the search-and-connect flow as not yet available; use the custom connector route.
- Redacted content is invisible to the connector. Where Compliance Check redaction is enabled, BrightHire removes the content from transcripts and search results, so an MCP answer will be missing it too. The feature is not retroactive and covers English transcripts only.
- Anthropic's verification is not a security audit. Anthropic states that the directory tier reflects the level of review a connector received, not how it works.
Frequently asked questions
Is the BrightHire MCP connector really read only?
Yes, and this one holds up under checking. BrightHire's own help centre documents exactly the same eleven tools Anthropic's directory lists, every one of them a retrieval verb, and its OAuth descriptors advertise a single scope named mcp:read.all. We found no vendor page, release note or scope describing an MCP action that changes BrightHire data.
What does get_transcripts actually return?
Full verbatim interview transcripts. BrightHire describes the tool as pulling complete transcripts plus user-created call notes for up to twenty interviews at a time, and warns that the payload is large. It recommends reaching for it only when exact wording matters, and suggests get_call_ai_notes as the lighter alternative for question-and-answer summaries.
Can the connector see interviews I do not have access to?
No. BrightHire states that MCP access follows the same permission model as the rest of the product, and that users can only reach data they already have permission to view. Visibility in BrightHire is scoped by hiring team, which the platform pulls from your ATS per position, and by department-level hiring team groups.
How does BrightHire handle candidate consent for recording?
BrightHire publishes an opt-in consent link that candidates verify by email before any recording happens. Where the link is detected on an imported calendar invite, the interview shows as awaiting candidate consent and BrightHire will not record until the candidate agrees. Consent can also be collected through the ATS application flow, and candidates may update their decision at any time.
Does BrightHire train AI models on our interview data?
No. BrightHire states on its responsible-AI page that it does not use customer data to train AI models, and its compliance page repeats the commitment as privacy-safe AI. BrightHire also describes Zero Data Retention arrangements with its AI sub-processors. Those statements cover BrightHire's own pipeline, not what your MCP client does with the answers.
Does the BrightHire MCP connector include Screen data?
No. BrightHire states plainly in its MCP documentation that the connector does not currently provide access to BrightHire Screen product data. Screen is BrightHire's AI interviewer product, so any AI-conducted screening interview sits outside the eleven tools on this page. Everything the connector reaches comes from human interviews captured in BrightHire.
Sources
- BrightHire MCP documentation — https://help.brighthire.ai/en/articles/14808753-brighthire-mcp (fetched 2026-08-19, HTTP 200, no redirect) · retrieved 2026-08-19
- BrightHire opt-in consent link — https://help.brighthire.ai/en/articles/15297790-opt-in-consent-link (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire Workday opt-in consent — https://help.brighthire.ai/en/articles/12814295-workday-opt-in-consent (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire Compliance Check — https://help.brighthire.ai/en/articles/8732434-compliance-check (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire hiring team groups — https://help.brighthire.ai/en/articles/10306947-hiring-team-groups (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire Advanced Insights permissions — https://help.brighthire.ai/en/articles/13130526-advanced-insights-permissions (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire release notes — https://help.brighthire.ai/en/articles/14069196-release-notes (fetched 2026-08-19; contains no MCP mention) · retrieved 2026-08-19
- BrightHire compliance and security — https://brighthire.com/compliance-and-security/ (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire responsible AI in hiring — https://brighthire.com/ai-in-hiring/ (fetched 2026-08-19) · retrieved 2026-08-19
- BrightHire privacy policy — https://brighthire.com/privacy-policy/ (fetched 2026-08-19) · retrieved 2026-08-19
- Live unauthenticated probe of
https://app.brighthire.ai/mcp/v1and both.well-knowndescriptor families — 2026-08-19 · retrieved 2026-08-19 - Anthropic connector directory — https://claude.ai/directory/024923b0-b122-4eb0-bb25-c3887bdaa9b9 (snapshot 2026-08-16) · retrieved 2026-08-16
Use in Agentman
Connect once and your agents call these tools on their own — on a schedule, in a workflow, with nobody at the keyboard.
Open in Agentman StudioServer Info
- Category
- Productivity
- Developer
- BrightHire
- Tools
- 11
- Domain
- app.brighthire.ai
Using Claude Desktop or another MCP client? Setup docs — the connection URL above works anywhere.