Support

Contact

The fastest path is the support widget in the bottom-right corner of the data viewer. It opens a messenger so replies are waiting for you when you come back.

You can also email us at support@viewparquet.com with a description of the problem, your browser, and file type (Parquet, GeoParquet, CSV, or JSON). Please do not attach sensitive datasets unless you have permission to share them.

We aim to acknowledge account, billing, privacy, and service-access requests within two business days. This target is not a paid support SLA. Include the account email and a safe description, but never send passwords, provider keys, presigned URLs, or full datasets.

Cloud files and projects

• Add Cloud copy fails: confirm you are signed into the intended account, the project still exists, and the file fits the plan's storage and per-file limits.
• Device and Cloud badges: these are independent copies. Removing a device copy does not remove S3, and removing S3 does not delete a saved device file.
• Move versus copy: Move changes project membership; Copy creates an independent identity and can consume additional Cloud storage.
• Missing Cloud object: add the source file to Cloud again. The query API will reject stale metadata before charging allowance.

Cloud queries

Cloud queries run as isolated, scale-to-zero DuckDB jobs and can take longer than device queries because a worker must start first. The job page shows starting, running, terminal status, runtime, and actual allowance used. Validation and launch failures are not charged; work that has started can consume its billable runtime even when cancelled or unsuccessful.

• Use device execution when every attached file is on this device.
• Use Cloud execution when any required attachment is Cloud-only or exceeds practical device resources.
• Result downloads expire after seven days; export anything you need to retain.
• Two jobs can run concurrently per account, and each job has a 15-minute hard limit.

Billing, cancellation, and credits

Polar is the Merchant of Record for self-serve purchases. The amount and currency in Polar checkout are authoritative. Manage renewal, payment methods, cancellation, and invoices from Settings → Usage & billing.

• Upgrades normally apply immediately with proration; downgrades apply at the next renewal.
• Credit top-ups are USD-only until an explicit local-currency policy is introduced.
• If payment succeeded but credits are missing, reopen Usage & billing to reconcile the order, then contact support with the Polar order ID if it remains unresolved.
• Sign in with the original or linked method if Polar reports an existing subscription under the same email.

S3 & remote URLs

Open Parquet and other formats from HTTPS links, public s3:// objects, or private buckets in the viewer via Open from S3 / URL (or /viewer?open=remote).

• Public URL: paste HTTPS or a presigned link; no credentials needed.
• Private bucket: use the Credentials tab with AWS, Cloudflare R2, GCS (HMAC), or MinIO presets. Keys sign requests in-browser only.
• 403 / CORS errors: check bucket policy and IAM permissions; for browser HTTPS URLs the bucket may need CORS on GET.

Step-by-step walkthrough: How to open Parquet from S3 privately.

AI assistant

The AI assistant is optional. A ViewParquet Cloud plan includes managed AI with no provider account or API key. Advanced users can instead add their own OpenRouter or supported direct-provider key under Settings → AI → Connection.

• Managed Cloud AI: sign in and choose a Cloud plan. DeepSeek V4 Flash is the default, and the chat model picker can search the live tool-capable OpenRouter catalogue. Provider credentials remain on the ViewParquet server.
• BYOK setup: create a key at openrouter.ai/keys and paste it in Settings → AI → Connection. Pick a model in the chat panel or Models settings.
• Data sharing: control what the assistant sees in Settings → AI → Privacy (column headers, types, metadata, sample values). Chat and SQL helper can use different presets.
• SQL Ask AI / Fix: use ⌘K or Ctrl+K in the SQL editor. It follows your selected BYOK model and your data-sharing settings.
• Usage and billing: AI usage and charges use your plan allowance in managed mode and your provider account in BYOK mode. Current managed allowance usage appears in Settings → AI.

For what leaves your browser when AI is enabled, see our Privacy Policy.

Already have a browser-capable AI agent? Read How to use AI with Parquet to let it operate the local viewer without buying or configuring ViewParquet AI.

Account data and privacy requests

Account settings can download a compressed JSON export containing account, project, conversation, usage, and Cloud-query records. You can delete individual Cloud copies and conversations or request account deletion. Confirmed account deletion immediately revokes a linked ViewParquet subscription and ends account access before Cloud data is removed. Temporary storage-provider failures are retried by a private deletion job. Contact us if cancellation cannot be confirmed or the privacy request needs to be handled separately.

Use the subject ViewParquet Privacy Request for access, correction, portability, objection, or erasure requests. Security vulnerabilities should follow the private reporting guidance on our Security page.

Resources

• Open Parquet from S3 privately : HTTPS, presigned URLs, and private buckets
• Parquet FAQ : how to open, view, query, and convert Parquet files
• Changelog : recent product updates
• Feature requests & feedback : vote on roadmap items. Posts are public, so do not include datasets, SQL, credentials, personal information, or other sensitive data.
• Blog : guides on Parquet, DuckDB, and data workflows
• DuckDB documentation : SQL reference for queries in the viewer

Company

ViewParquet is operated by Kaladev Digital Private Limited and built by Kaladev Digital. Follow updates on X (@antonyadhiban).