Privacy Policy

Scope

This policy covers the Unbuilt website, desktop application, account services, live model requests, and interview-library storage. The exact data involved depends on the features you choose to use and the material you choose to share or upload.

Information we handle

• Account and access information. We handle information used to sign you in and control access, such as your email address, account identifier, verification and plan state, and invitation or access-code records.
• Usage and operational information. We record service activity needed to operate and protect the product, including request and token usage, timestamps, access events, and limited technical diagnostics.
• Website analytics. When you visit unbuilt.build, we collect the page and tracked links you use, time, referring site, an anonymous browser visitor and session identifier, browser and operating-system family, your IP address, and approximate country, region, city, and time zone. We do not store the full browser user-agent string, and website analytics are kept separate from interview data.
• Live assistant inputs. When you use the assistant, selected shared-screen images, recognized problem text, conversation context, and audio portions may be sent for model processing so the product can provide guidance.
• Saved interview records. While you use the product with a verified signed-in account, the app may automatically queue and synchronize a structured archive containing transcripts, questions, answers, replies, feedback, and related usage details. A failed or pending archive upload may be retried when that account later signs in with verified access. The structured interview archive itself does not include screenshot or audio bytes.
• Optional uploaded audio. Where an existing product feature separately uploads an audio file, that audio may be stored apart from the structured interview record. Audio is not included merely because a structured archive is synchronized.

Screen and audio boundaries

Unbuilt processes selected screen images and audio through the assistant features you use, including by sending that material to the model service for live guidance. The global Pause control stops new live audio and screen inputs. Pause screen stops new screen analysis while call audio and transcript handling continue. The structured interview archive excludes raw screenshots and audio bytes; where an existing feature separately uploads audio, that audio is stored separately as described above.

Safe diagnostics are designed to omit interview content. Optional debug logging, when enabled, can contain transcript text, so debug logs should be handled accordingly.

How we use information

We use this information to authenticate accounts, enforce invitation and usage limits, provide live guidance, synchronize interview records, generate interview review features, troubleshoot the service, prevent misuse, and understand service usage. We do not describe company marks shown on the site as customer relationships; they are presented as aspirational interview targets.

Website analytics help us understand visits and clicks, including broad geographic and technical patterns. Raw visitor logs, including IP addresses and approximate locations, are visible only to Unbuilt administrators.

Authorized Unbuilt operators and administrators may access stored account information, usage records, and structured interview records when operating, supporting, securing, or reviewing the service.

Cloud and model providers

We use Google Cloud services, including Vertex AI, as infrastructure and processing providers. Depending on the feature, those providers may receive account-related service data, live model inputs, usage measurements, structured interview records, or optional audio in order to host storage, authenticate requests, operate the service, or generate model output.

These providers process information through their own infrastructure on our behalf. Their systems, security controls, logs, and retention practices are not the same as Unbuilt’s application-level storage, so this policy does not promise that provider-side processing creates no records or that provider systems delete data on an application-defined schedule.

Storage, retention, and deletion

Saved interviews can exist locally and, for signed-in accounts, in cloud storage under the account’s identifier. Local archives are encrypted only when compatible operating-system secure storage is available; otherwise session recovery remains in memory instead of being written as a recoverable local archive.

Unbuilt does not currently promise a fixed retention deadline for account information, usage records, uploaded interview records, or optional audio. An individual interview’s delete action removes the local archive and requests deletion of the matching cloud interview record and separately stored audio. A cloud deletion can be delayed by connectivity, authentication, or storage failures; the app reports that result and can retry the recorded deletion intent during a later eligible session.

Website analytics records, including IP addresses and approximate locations, are retained indefinitely. They are visible only to Unbuilt administrators. To request deletion of visitor analytics associated with your IP address, email contact@unbuilt.build.

That action covers the interview objects the app directly controls. It is not a promise of immediate deletion from provider systems outside those objects, such as provider-operated logs or backups. For privacy or deletion questions beyond the in-product interview action, contact us.

Security and choices

Account-scoped authorization limits ordinary users to their own interview storage, and the product uses the available platform and cloud security controls. No online system can guarantee absolute security. You can pause live observation, delete individual saved interviews through the product, or contact us with questions. Signed-in archive synchronization may occur automatically as described above.

Job applications

If you apply for a job through unbuilt.build/careers, we keep what you send: your name, email address, the role, the link and your message, with the time it arrived. We also keep where the application stands (new, contacted or closed) and a keyed fingerprint of the network address it came from, which we use only to limit repeated submissions.

Applications are stored in our Google Cloud project, with account data, and only the Unbuilt team can read them. When an application arrives, an email with the role, your name, email address, link and the first 300 characters of your message is sent to contact@unbuilt.build.

We keep an application to consider you for a role and to reply. Email contact@unbuilt.build to have it deleted, and we will delete it.

Changes to this policy

We may update this policy as the product or its data flows change. The date above identifies the current version. Material changes will be reflected on this page before or when the changed practice takes effect where practical.