Tempra Subprocessor and Technology Provider List

Effective date: October 2, 2026

Last updated: October 2, 2026

Tempra operator: Jason Chan, doing business as Tempra

This page identifies third parties that support Tempra's implemented account, billing, hosting, job-application automation, and website functions. It supplements the Privacy Policy, which explains the information Tempra processes, why it is processed, and users' choices and rights.

The term “provider” is used broadly here. A provider's legal role can depend on the service and activity. Some may process information on Tempra's instructions, while others may act independently for functions such as payment compliance, fraud prevention, or account administration. Inclusion on this page does not by itself establish that an entity is a “processor,” “service provider,” “contractor,” or “third party” under a particular law.

Provider participation can depend on the feature, application page, subscription tier, availability, and fallback route. One application can involve more than one AI provider. Information visible in a resume, profile, application page, screenshot, or browser recording can include personal information and sensitive information even when the provider's primary function is technical.

Who can see what.

• Create an account or pay: Clerk; Stripe through Clerk Billing.
• Save a profile, resume, or academic transcript: Neon (database and private files); DigitalOcean (while the API or a job is running).
• Run a supported Workday, Ashby, or Greenhouse application: Anchor (hosted browser and requested video); DeepSeek, Google Gemini, Alibaba Qwen, and/or OpenAI depending on the page and fallback.
• Appear on a recorded session: Anchor; authorized Tempra operators reviewing a recording.
• Sit in the job queue: Upstash Redis (job id and dispatch metadata, not the whole profile).
• Visit the marketing site: Google Fonts and any other remote website assets actually loaded.

Employers and their recruiting platforms, including Workday, Ashby, and Greenhouse, are not Tempra subprocessors merely because Tempra operates their application pages. They receive whatever is typed, uploaded, saved, or submitted on their site.

Providers supporting the customer service

Clerk

• Function: Authentication, account management, and subscription administration.
• Information involved: Account identifiers, authentication and session information, account status, subscription information, and billing-related metadata.
• Purpose: Verify account access, maintain account state, present account-management controls, and administer plan entitlements.

Stripe, through Clerk Billing

• Function: Payment processing for Clerk Billing.
• Information involved: Payment and customer information needed for checkout and billing, transaction status, plan information, and related fraud or compliance data. Tempra's application database is not intended to store full payment-card numbers or card security codes.
• Purpose: Process payments, maintain subscription and transaction status, prevent payment fraud, and meet payment-related legal requirements.

Neon

• Function: Hosted PostgreSQL database and private S3-compatible object storage.
• Information involved: Account and profile information; career, eligibility, and optional sensitive information; resume and academic-transcript files and metadata; application snapshots and history; job, credit, subscription, audit, and usage records; diagnostic screenshots, pipeline artifacts, and associated identifiers.
• Purpose: Persist account and application state, store private uploaded and generated files, enforce ownership and usage rules, support application history, and retain operational and diagnostic records.

DigitalOcean

• Function: Hosting and compute for Tempra's API, automation workers, and related processes.
• Information involved: Information processed by the API and automation jobs, including account and application identifiers, profile, resume, and academic-transcript information materialized for a job, employer-page content, application answers, diagnostics, request metadata, and temporary execution files.
• Purpose: Run the customer-facing API and job-application automation and support service operation, security, and troubleshooting.

Upstash

• Function: Redis infrastructure used for Taskiq job dispatch.
• Information involved: Job identifiers and operational dispatch metadata. Queue messages are not intended to contain the user's whole profile or resume; Neon remains the durable source for job and credit state.
• Purpose: Dispatch authorized jobs to workers and coordinate execution.

Anchor Browser

• Function: Hosted employer-browser sessions and continuous visual session recording.
• Information involved: Employer and applicant-tracking-system URLs and pages; screen content; mouse, keyboard, and explicitly pasted clipboard input; uploaded resume and academic-transcript content; browser cookies and session state; login activity; profile and application information displayed or entered in the session; and continuous MP4 recordings of managed sessions.
• Purpose: Open and operate employer application pages, support user sign-in and review, perform authorized automation, diagnose failures, investigate support or safety issues, and improve service reliability.

DeepSeek

• Function: AI generation of profile-based job-application answers. DeepSeek is a China-based provider.
• Information involved: Application questions, choices, page context, prompts and outputs, and the saved application profile supplied to the question-answering path, including filled sensitive fields.
• Purpose: Interpret application questions and generate or select proposed responses.

Google, including Gemini

• Function: AI question answering, narrative generation, application-page and image interpretation, Autopilot form-availability and submission-result screenshot checks, and fallback processing on configured paths.
• Information involved: Application questions and options, prompts and outputs, saved profile context supplied to answering paths, screenshots, page content and structure, and information already visible or entered on an application page.
• Purpose: Extract application fields, interpret page content, generate or select proposed answers, assess the visible Autopilot form and submission result, and support automation when the configured path uses Gemini.

Alibaba Cloud, including Qwen

• Function: AI image-based form extraction and control location on selected processing paths.
• Information involved: Application screenshots and other image inputs, page questions and controls, extracted form structure, prompts, outputs, and information already visible or entered on the page.
• Purpose: Detect, classify, and locate application fields and controls so the automation can interact with the selected employer page.

OpenAI

• Function: AI interpretation of employer-page text or structure and identification of page content and controls.
• Information involved: Employer-page content, extracted text or structure, prompts and outputs, control information, and personal information already present in that content.
• Purpose: Help identify and interact with application-page content and controls on configured automation paths.

Website asset recipients

The current marketing website requests assets from Google Fonts and third-party image hosts identified in the source and Privacy Policy, including Framer, Massive, and Google-hosted image or favicon services. Those requests can disclose network and request information such as an IP address, browser information, requested resource, request time, and, depending on browser and request behavior, a referring page or site address.

Jobs-board company icons also load from employer, Ashby, Greenhouse, and Google favicon hosts. These hosts receive network and request information needed to serve the icon. The board requests icons without a referring-page header.

These recipients are listed for transparency and are not necessarily subprocessors acting solely on Tempra's instructions.

Providers that are not covered by this list

Employers, applicant-tracking systems, and services embedded in employer application pages receive information when you or Tempra enter, upload, save, or submit it on an application you select. They determine their own recruiting and website practices and ordinarily are not Tempra subprocessors. Their privacy notices govern their independent processing.

Professional advisers, auditors, authorities, courts, transaction counterparties, and parties that receive information at a user's specific direction are also not recurring service providers merely because a legally permitted disclosure can occur. The Privacy Policy explains those disclosure categories.

Experimental, archived, local-development, and disabled fallback integrations are not included as active customer providers.

Data-use conditions for providers

• Current answering routes can send an AI provider the user's whole saved application profile, including filled fields unrelated to the immediate question, and image routes can send screenshots containing incidental page content.
• Tempra does not sell or license customer profiles, resumes, application content, browser recordings, or diagnostics to AI labs, data brokers, or advertisers for compensation, model training, or targeted advertising.
• AI providers ingest and process the information Tempra sends to return outputs. Depending on a provider's terms, service tier, endpoint, and account settings, the provider may retain, permit human review of, use to improve services or models, or train on that information. Tempra cannot guarantee that every active or fallback provider excludes those uses.

Using personal information to return an AI output is distinct from training or fine-tuning a model. Providers may still retain or permit limited access to request content for service delivery, security, abuse prevention, or legal compliance.

Changes to this list

Tempra will update this page and its “Last updated” date when it adds, replaces, or materially changes a provider that processes customer personal information. When required by law, Tempra will provide notice of a material change, such as an in-service notice or email, and will obtain additional consent when required. A new provider does not authorize an undisclosed incompatible use of information.

Contact

For questions about Tempra's providers or this list, contact jason@usetempra.com.