How we rank
Every tool is scored 0–10 on the same six criteria. The final score is the weighted average, to one decimal.
| Criterion | Weight | What a 10 looks like |
|---|---|---|
| Job fit | 25% | Clearly does the category job end to end |
| Output quality | 25% | Results usable without heavy fixes on our standard test task |
| Ease of start | 15% | Value within 10 minutes, no sales call |
| Pricing & value | 15% | Transparent pricing, a usable free tier or fair per-use cost |
| Reliability & control | 10% | Consistent runs, user approval/undo, clear errors |
| Momentum & openness | 10% | Active releases, docs, API/CLI, community |
Standard test tasks
Each category uses one fixed test, for example "build a todo app with auth" for vibe coding or a 10-second product clip for video, so tools are compared like for like.
Tested vs. untested
Tools we haven't tested yet carry a Tested: no label and a provisional score. Their verdicts are based only on the linked public sources, and they can't rank #1.
Paid placements & affiliate disclosure
- Paying never changes a tool's score or rank.
- Paid placements appear in a separate row labelled Featured.
- We have no affiliate links today. If we add any, they'll be marked on each card and won't affect rankings.
- Every card links to its source and shows when we last checked it.
AgentPick Weekly
The week's new agent tools with a verdict on each, every week. Free. We'll email you a link to confirm, and you can unsubscribe any time. See our Privacy Policy.