Privacy Policy
Effective and last updated: October 20, 2026 · Version 2026-10-20
This Privacy Policy explains how PT Netra Runtime Indonesia ("Netra", "we", "us", or "our") processes personal data when you use the Netra Runtime website, platform, APIs, and related services (the "Service"). Netra acts as controller for its account and operational purposes and processes customer-controlled request content according to the applicable customer instructions and agreement.
Our registered office is Mula Kantor Indonesia Lt. 2, Haery 1 Building, Jl. Kemang Selatan No. 151, RT 004/RW 004, Cilandak Timur, Pasar Minggu, Jakarta Selatan, DKI Jakarta 12560, Indonesia. For privacy requests, contact rifky@netraruntime.com. The Bahasa Indonesia version is the primary version for Indonesian users.
1. Scope and age restriction
The Service is currently offered in Indonesia to adults and organizations. It is not intended for children, and users must be at least 18. We do not knowingly permit a child to create an account or intentionally collect a child's personal data. Contact us if you believe a child has provided data so we can investigate and delete it where required.
2. Data we collect
- Account and identity data: email address; name, avatar, and identity-provider identifiers when you use Google or GitHub sign-in; and the time and version of your legal acceptance and adult confirmation.
- Profile and billing data: individual or business account type, name, company details, phone number, address, optional tax identifier, billing email, invoices, and transaction references.
- Payment data: payment provider, amount, currency, method category, status, and provider reference. Polar or DOKU receives payment credentials directly; Netra does not store full card or bank-account numbers.
- API and operational metadata: organization and API-key identifiers, model, token counts, cached tokens, charge, timestamps, request ID, latency, status, errors, IP-derived security logs, and content-free traces needed for billing, reliability, abuse prevention, and support.
- AI request content: prompts, messages, source code, inference files, tool definitions, tool-call arguments, tool results, and generated responses are processed to answer a request. Inference content recording is off by default. Organization administrators may explicitly enable API and exported trace content recording for Observe and configured evaluations. Saved content: Each new Playground chat stays temporary until you select Save. Saving one chat does not automatically save later chats. Saved chats are associated with the workspace selected when you save. Uploaded files, saved application prompts and intentionally saved histories are stored to provide those features and follow the maximum 30-day content lifetime described below.
- Device and access data: IP address, browser or client type, operating system, referring page, and access timestamps received by Netra and its infrastructure providers when serving or securing the Service.
We do not sell personal data or use inference content for advertising. Optional advertising measurement on our website is subject to its consent controls and notice; usage, payment and security processing are separate from prompt content.
3. How and why we use data
We process data to:
- create and authenticate accounts and record your agreement to applicable terms;
- provide inference, API, billing, observability, support, and account features;
- measure usage, deduct credits, issue invoices, reconcile payments, and meet tax and accounting obligations;
- route a request to the selected model infrastructure and return the result;
- provide request recording and evaluations only when an organization administrator enables recording;
- protect accounts, investigate abuse or fraud, debug failures, and maintain service reliability;
- respond to rights requests, legal process, and regulatory obligations; and
- communicate material service, security, legal, or transaction information.
Depending on the context, our basis is performance of our contract, compliance with law, your consent for optional processing, and our legitimate interest in operating and securing the Service. Where consent is the basis, you may withdraw it without affecting earlier lawful processing.
4. Inference data
When you use Netra's covered inference APIs, Playground or Free OCR, we process the content you send to provide the requested result. This includes prompts, code, files, images, tool calls, tool results and generated responses.
We do not use your inputs or outputs to train models, fine-tune models, advertise, or improve our models or services. Saving content or requesting support does not give us permission to use it for those purposes.
Content recording is off by default. We do not persistently store inference content in logs or backups unless you explicitly enable recording, save content, or provide it for support. In-memory inference caching is separate and is explained in our Inference Data Handling Policy.
If you choose to save or record content, we delete it within 30 days of its original capture or Save, including recoverable backup copies. Editing or viewing it does not restart that period. Support access is limited to authorised staff handling a support request you initiate.
We retain account, billing and content-free operational information separately for purposes such as authentication, payments, security and service reliability. These records may include request identifiers, model names, token counts, timestamps and status codes.
Native inference runs on Netra-operated infrastructure in Indonesia. Eligible requests may use another inference provider when the primary service cannot fulfil them. Fallback may involve processing outside Indonesia and must meet the applicable data-handling commitments. Native-only routing is available by arrangement once Netra confirms its activation.
For privacy questions or deletion requests, contact rifky@netraruntime.com, with farrel@netraruntime.com as the backup contact.
6. International transfers and data location
The platform database is operated in Jakarta, Indonesia. Native AMD inference is operated in Indonesia. Fallback to another inference provider subject to this policy's content-handling requirements may occur when the primary inference server is unavailable, returns an error, or cannot process a request. Such fallback processing may take place outside Indonesia. These providers, their support teams and other recipients may process data outside Indonesia. A global routing label is not a guarantee of physical processing location. We assess vendors and use contractual, organizational, and technical safeguards appropriate to the transfer and applicable Indonesian personal-data law. We maintain records of the locations used to manage, process, and store the electronic system and personal data and will provide legally required information through our regulatory filings and notices.
7. Retention
The Inference Data Handling Policy governs inference-content retention. The following retention rules apply.
| Data | Retention rule |
|---|---|
| Inference content with storage disabled (default), including prompts, code, tool results and outputs | No persistent content logs or backups. Native shared RAM/GPU caching lasts up to seven days from entry creation without reset; per-workspace opt-out blocks shared-cache reads/new writes, while existing entries retain their original expiry. See the Inference Data Handling Policy |
| API/trace recordings and customer-configured evaluation content explicitly enabled by an authorised administrator | Delete within 30 days of first capture, or earlier on deletion; no use for Netra's own training or improvement |
| Each Playground chat explicitly saved by you, associated inference files, saved prompt material and tool results | Delete within 30 days of initial Save/capture, or earlier on deletion. Each new chat requires its own Save; edits and replays do not extend the original deadline |
| Inference content explicitly provided for support | Only for that support request, with the same maximum 30-day retention |
| Account and profile data | While active, then delete or de-identify on closure unless a specific lawful retention obligation applies |
| Content-free usage, credit, invoice, payment, fraud, tax and accounting records | For the applicable legal retention period and as necessary for reconciliation, fraud prevention and legal claims; these records must not contain inference content |
| Versioned legal-acceptance evidence | Pseudonymous evidence retained as reasonably necessary for legal claims |
| Content-free security and access logs | Limited to the period needed for security and incident response, separately from inference content |
Disabling recording stops new captures; it does not extend existing deadlines. Inference-content deletion must cover active storage and recoverable backup copies within the promised period. Backup recovery must not reintroduce expired or deleted content. An access restriction alone is not deletion. If binding law requires preservation, we limit it to the required content, purpose and period.
8. Security
We use HTTPS for public access, tenant-scoped access controls, hashed API-key secrets, scoped service credentials and audit controls. Offsite recovery archives are encrypted by Netra before upload. Netra manages these encryption and recovery keys; this is not end-to-end encryption. Encryption does not authorise retention of inference content outside the handling rules described here or extend any deletion deadline. Support access to customer-saved content is limited to Alifais Farrel Ramdhani and Rifky Bujana Bisri, only at the customer's request and only as necessary for that case; access does not extend the original deletion deadline. Privileged infrastructure access remains restricted to authorized operators and subject to operational controls. No system is completely secure. Protect your sign-in email and API keys and promptly revoke exposed keys.
9. Your rights
Subject to Indonesia's Personal Data Protection Law (Law No. 27 of 2022) and other applicable law, you may request information about processing; access or a copy; correction; deletion or destruction; withdrawal of consent; restriction or objection; data portability where applicable; and compensation or other remedies for unlawful processing. Some requests may be limited where retention or processing is legally required.
Email rifky@netraruntime.com to exercise a right. Describe the request and the account email. We may verify identity before acting and will respond within the period required by law.
10. Automated processing
Netra uses automation to route inference, meter usage, deduct credits, detect technical failures, and apply security limits. Netra does not use the Service to make a decision on your behalf that produces legal or similarly significant effects. Model Outputs are generated probabilistically and require human review.
11. Changes
We may update this Policy as the Service, providers, or law changes. We will change the date and version above and provide additional notice for material changes. Where a change requires renewed agreement or consent, we will request it before the affected processing continues.
12. Contact
PT Netra Runtime Indonesia
Mula Kantor Indonesia Lt. 2, Haery 1 Building; Jl. Kemang Selatan No. 151, RT 004/RW 004; Cilandak Timur, Pasar Minggu, Jakarta Selatan 12560, Indonesia.
Privacy contact: rifky@netraruntime.com