Editorial Standards
How we review and rate AI tools — and what we will never do.
Independence First
PromptBulletin editorial scores and rankings are never influenced by advertising, affiliate commissions, or tool vendor relationships. Sponsored placements are clearly labeled with a "Sponsored" badge and do not appear in organic ranked results. We decline vendor requests to review unpublished tools under NDA arrangements that could compromise our ability to disclose limitations.
Our 12-Point Scoring Rubric
Every tool reviewed on PromptBulletin is evaluated across twelve criteria, each scored on a 1–10 scale. The final Editor Score is a weighted average that emphasises performance and value, the two factors our readers care most about.
Performance
Does the tool deliver on its core promise? We test with real tasks, not demos.
Value for Money
Is the pricing fair relative to what competitors offer at the same tier?
Ease of Use
How quickly can a new user get meaningful output? We time first-run onboarding.
Output Quality
For generative tools, we evaluate accuracy, coherence, and practical usefulness of outputs.
Feature Depth
Does the tool have the controls and customization power that advanced users need?
Documentation
Is the help center, API docs, and support material complete and easy to navigate?
Reliability
Uptime, latency, and stability under normal load over a multi-week testing period.
Integrations
How well does it connect with the rest of a modern workflow — APIs, webhooks, popular apps?
Privacy & Security
Data retention policies, SOC2/GDPR compliance status, and how user data is used for training.
Support Quality
Responsiveness and helpfulness of customer support across free and paid tiers.
Update Cadence
Is the team actively shipping? Dead tools score lower regardless of current feature set.
Community & Reputation
User sentiment from independent sources — Reddit, G2, Trustpilot — cross-referenced against our own testing.
Testing Process
Editors test each tool hands-on using a standardised test suite before publishing a review. For AI generative tools, we run a fixed set of prompts and compare outputs across multiple sessions. Testing typically takes 1–3 weeks. We re-review tools annually or when significant pricing or feature changes occur.
Corrections Policy
We correct factual errors promptly. If you spot an error in a review — incorrect pricing, outdated features, or a factual mistake — please contact us. We note significant corrections at the bottom of the article with the date of the update.
Affiliate Relationships
Some tool links on PromptBulletin are affiliate links. If you click through and purchase, we may earn a commission. This never affects our ratings. See our full Disclosures page for details.