Last updated: October 30, 2026
A subprocessor is a third-party service provider that Cohera AI ("Cohera") uses to process personal information on behalf of the employers who use the platform — for example, to host the service, store interview recordings, transcribe speech, or run the AI evaluation. Cohera engages each subprocessor under a written contract that restricts its use of personal information to providing the service to Cohera and requires appropriate security and confidentiality commitments.
This page is part of our Data Processing Addendum and supplements section 5.3 of our Privacy Policy.
1. Current subprocessors
| Name | Purpose | Data categories | Region |
|---|---|---|---|
| OpenAI | Realtime voice interviews and speech transcription; language-model evaluation of interviews and assessments; the model behind the Codex coding tool candidates use in AI-Assisted assessments | Interview audio, transcripts, code and design artefacts, evaluation outputs; candidates' prompts to the coding tool and the code it reads | United States |
| ElevenLabs | Speech synthesis: turns the AI interviewer's written replies into its voice during realtime voice sessions | The text of the AI interviewer's spoken replies, which can include the candidate's first name and short quotes of their answers; no candidate audio, transcript or recording. Request history is disabled | United States |
| Anthropic | Language-model authoring and grading of assessments; the model behind the Claude Code coding tool candidates use in AI-Assisted assessments; converting an employer's own repository into an assessment | Transcripts, code, submissions, rubrics; candidates' prompts to the coding tool and the code it reads; employer-supplied repository code | United States |
| CoreSignal | Candidate sourcing data provider | Public professional profile data | European Union / United States |
| FullEnrich (FullEnrich Corp) | Personal email address and, only on a recruiter's request, personal mobile phone number discovery for sourced candidates | Candidate name, current employer and its domain, public LinkedIn profile URL; returned personal email address with a deliverability status and mobile phone number with its line type and verification status | United States |
| Google Cloud / Firebase | Authentication, Firestore database, Cloud Storage for recordings and uploaded files | Account data, role data, interview recordings, uploaded files | United States |
| Supabase | PostgreSQL database | Candidates, bookings, evaluations, billing records | United States |
| Render | Backend API and worker hosting | All data processed by the API | United States (Ohio) |
| Vercel | Frontend hosting and edge network; isolated cloud workspaces (Vercel Sandbox) for AI-Assisted assessments, deleted when the session ends | Request metadata, page content; in assessment workspaces, the candidate's code, terminal activity and the employer's task code | United States |
| Sentry | Error monitoring | Technical error data, which may include user identifiers | United States |
| PostHog | Website and product analytics | Page views, click events, referrer and campaign parameters, account identifier for signed-in users | United States |
| Stripe | Payment processing | Customer billing details, payment status | United States |
| Cloudflare | DNS and TURN media relay for WebRTC interviews | Connection metadata; relayed (encrypted) media streams | Global |
| Google Workspace | Outbound and inbound email | Email addresses and message content | United States / Canada |
2. Changes to this list
We update this page when we add or replace a subprocessor. Customers who would like to be notified of changes can ask by emailing privacy@coheraai.com; we will send notice to the address you nominate before the new subprocessor processes your candidates' personal information, as described in the DPA.
3. Questions
Email privacy@coheraai.com.