Research investigation R0914 / claim audit

Bolt Forge’s 50× Trade: What a Builder Consents to Donate for Open-Model Training

The supplied record describes a substantial opt-in bargain, including broad Forge Content classes, mode-specific consent, some exclusions, and prospective withdrawal controls. It does not fully specify the transformation procedure, secret-removal limits, retention of training copies, the non-public Bolt-Arcee agreement, or deletion from completed training and delivered datasets.

Current public editionv1Sep 14, 2026
Verified observations
20

16 measured fields

Supported claims
14

14 material findings

Cited sources
8

8 primary or authoritative

Research score
86

Automated topic and evidence score

Interactive figureBolt Forge’s 50× Trade: What a Builder Consents...
CSV JSON
Data status15 verified records across 1 period

Snapshot only. There is not enough history to claim a trend yet.

Verified observationHover or focus any mark for exact valuesLast updated Sep 14, 2026

Version ledger

Frozen public editions

Each edition preserves the records, method, sources, and downloads available at publication time.

  1. v1 / latestSep 14, 202620 records / 8 sources

    Initial public snapshot with 20 records and 8 cited sources.

Coverage note

This synthesis is limited to the supplied official-page evidence packet and its stated cutoff. It does not establish what appears in private agreements, authenticated controls, recipient systems, or later revisions.

Dataset ID
spd:bolt-forge-s-50-trade-what-exactly-does-a-builder-consent-to-donate-for-open-model-training-93b36da9
Stable URL
/research/bolt-forge-s-50-trade-what-exactly-does-a-builder-consent-to-donate-for-open-model-training-93b36da9
Version
v1
Coverage
2026-09-14
Records
20
Fields
7
Updated

Read the data

The records behind the figure

CSV JSON
Bolt Forge’s 50× Trade: What a Builder Consents to Donate for Open-Model Training data records
EntityMetricValueUnitObservedSourceTransform
StackBlitzActive-account retentionPersonal information and user content retained while account remains active2026-09-14https://stackblitz.com/privacy-policy
StackBlitzAnonymization disclosureRemove or transform reasonably identifying information and maintain de-identified form2026-09-14https://stackblitz.com/privacy-policy
StackBlitzAnonymization procedureUnspecified: no published algorithms, thresholds, test results, or error rates2026-09-14https://stackblitz.com/privacy-policy
Bolt ForgeData class disclosureAI Input, including prompts, code, files, and other submitted content2026-09-14https://stackblitz.com/terms-of-service
Bolt ForgeData class disclosureAI Output generated while Forge is enabled2026-09-14https://stackblitz.com/terms-of-service
Bolt ForgeData class disclosureError messages, correction and fix sequences, tool invocations, and edit histories2026-09-14https://stackblitz.com/terms-of-service
Bolt ForgeData class disclosureProject files, code, and configuration to the extent processed by Forge2026-09-14https://stackblitz.com/terms-of-service
Bolt ForgeData class disclosureAccount identifiers used to associate content with an account before de-identification2026-09-14https://stackblitz.com/privacy-policy
StackBlitzDeletion effectRemoves requested content from training corpora and datasets not yet delivered, subject to exceptions2026-09-14https://stackblitz.com/privacy-policy
Bolt ForgeDeletion from trained modelsNot unwound by stopping Forge, withdrawal, or deletion alone2026-09-14https://stackblitz.com/terms-of-service
EEA, UK, and Swiss accountsForge training-data treatmentForge available, but content is not used for model development or dataset licensing and consent is not requested2026-09-14https://stackblitz.com/terms-of-service
StackBlitzMaximum withdrawal implementation period15days2026-09-14https://stackblitz.com/terms-of-service
Arcee AINamed training recipientNamed for the first preview training run under a signed, non-public DPA2026-09-14https://bolt.new/blog/what-is-bolt-forge
Dataset licenseesRecipient restrictionsNo re-identification; onward disclosure only under equivalent restrictions2026-09-14https://stackblitz.com/privacy-policy
Bolt ForgeSecret-removal validationSeeded test data is claimed; protocol and results are not published2026-09-14https://bolt.new/blog/what-is-bolt-forge
Supplied official-page snapshotsSnapshot retrieval after fixed cutoffApproximately 3 minutes 20 secondsseconds2026-09-14https://stackblitz.com/privacy-policyComputed 2026-09-14T20:33:19.565Z minus fixed cutoff 2026-09-14T20:30:00.000Z.
StackBlitzVoluntary-closure deletion maximum30days2026-09-14https://stackblitz.com/privacy-policy

Measurement technique

How to read this report

  1. 01Build an evidence matrix with rows for data classes, consent, anonymization, secret removal, recipients, retention, withdrawal, deletion, derivative models, and exclusions.
  2. 02Classify each cell as expressly disclosed, partially disclosed, unspecified, or mixed across the supplied official materials.
  3. 03Separate launch-page statements from the StackBlitz terms and privacy policy, and distinguish ordinary account-content rules from Forge training-corpus, recipient-copy, and model-treatment rules.
  4. 04Treat the authenticated consent screen and the Bolt-Arcee DPA as unavailable rather than infer their contents.
  5. 05Preserve the fixed September 14, 2026, 20:30 UTC cutoff. The supplied snapshot-retrieval observation occurring after that cutoff is a collection-timing limitation, not new evidence of the documents’ contents.
Next report / 01AI Model Economics Index All research reports