How we score

Two scores, no overall number

Each tool has a Pricing score and a Features score, both out of 10 and built from the criteria below, using the tool’s official pages. We do not combine them into one number. A single score hides what it rewards: a cheap tool with a long feature list can outscore a better one. Read the two scores, and the reasons beside them, side by side.

How open, fair, and good value the price is next to similar tools. Out of 10.

Source: The tool's official pricing page, plus its FAQ and billing pages.

• Price transparency

  2 points

  Can you see every plan's price and limits without talking to sales?

  Full marks

  Every plan's price and limits are public.

  Zero

  Most tiers say “contact sales”.

• Entry access

  2 points

  How easily can you try it?

  Full marks

  A free plan with no time limit scores 2. A trial with no card scores 1.5. A trial that needs a card, or a cheap entry plan, scores 1.

  Zero

  Paid only, with no trial.

• Value per dollar

  3 points

  Next to peers in the same category, what does the entry paid plan cost for the limits it gives?

  Full marks

  25% or more below the category median price for comparable limits scores 3. Within 25% of the median scores 2. 25% or more above scores 0 to 1.

  Zero

  Well above peers for the same limits.

• Limit clarity

  2 points

  Are caps, credits, and overage prices stated?

  Full marks

  Caps, credit costs, and top-up prices are all stated.

  Zero

  Caps are hidden or vague.

• Billing fairness

  1 point

  Is billing flexible and fair?

  Full marks

  Monthly billing, a yearly discount, and rollover or refunds on failed work. Each missing item loses a share of the point.

  Zero

  Annual only, with no rollover and no refunds.

• The category median is the median entry paid price of the tools we list in that category. It is used only for this score, not for tags.
• A free sign-up that leads straight to a paywall is not a free plan.
• If a tool's own pages say a plan cannot be bought yet, we say so in an Availability note. It is a warning to the reader, and it does not change the points.

How much of the category's core job the tool covers, and how well that is shown. Out of 10.

Source: The tool's official product, docs, and feature pages.

• Coverage

  6 points

  How many of the 10 locked feature rows does it cover?

  Full marks

  Yes is 1, Partial is 0.5, No and Not listed are 0. Add the rows, divide by 10, and multiply by 6. This part is calculated, not judged.

  Zero

  No row is a Yes or Partial.

• Depth

  2 points

  Are the Yes rows backed by named capabilities, not vague claims?

  Full marks

  Each Yes names what it does, and any plan gating is stated. A feature only on the top plan counts as half.

  Zero

  Claims are vague or unverifiable.

• Differentiator

  1 point

  Does it offer something most peers lack?

  Full marks

  A named capability that most tools in the category do not have.

  Zero

  Nothing that peers do not also offer.

• Reliability signal

  1 point

  Can the claims be checked?

  Full marks

  Public docs, a changelog, or an API reference exist.

  Zero

  No docs and no changelog.

• Not listed is not No. It earns no points but is shown as “not listed”, so a tool is never said to lack something it may have.
• Scores compare tools within one category. A tool is never scored against a different category's feature rows.

Reading a 0–10 score

• 9–10Best in categoryHard to beat on this pillar.
• 7–8.9StrongClearly better than most peers.
• 5–6.9AverageIn line with most peers.
• 3–4.9WeakBehind most peers.
• 0–2.9PoorWell behind peers or mostly missing.

Why-switch tags

Tags such as Cheaper or Better for teams say why someone might switch to a tool. They are awarded by the rules below from official facts. They are not part of the 0–10 score, and the score never decides a tag.

• Cheaper

  Qualifies

  The category lists at least 6 tools, and the entry paid plan is among the 3 lowest entry prices in it, with comparable limits. A tool with no paid plan is not counted.

  Ranked by

  Highest value: lowest comparable entry price. A tie on price goes to the plan that includes more for it (more prompts, credits, seats, articles or exports for the same job). The Pricing score is only a last tiebreaker.

  Source

  The official pricing page.

• Better agent

  Other categories

  Qualifies

  Agent access (MCP, CLI, or API) is on the entry plan, and an agent can create or schedule, not just connect.

  Ranked by

  Highest value: most agent surfaces. A tie goes to the higher Features score.

  Source

  The official pricing page, docs, and product pages.

  AI Video

  Qualifies

  MCP, CLI or API access is on the entry paid plan and runs on that plan's credits, with no separate developer account or billing. An agent can generate or edit video, not just read.

  Ranked by

  Highest value: most agent surface types (MCP, CLI, API, SDK, plugin or skill, each counted once). A tie goes to the higher Features score, then to the higher Pricing score.

  Source

  The official pricing page, docs, and product pages.

  AI Meeting Notes

  Qualifies

  MCP, CLI or API access is on the entry paid plan and runs on that plan, with no separate developer account or billing. An agent can search meetings or create notes and action items, not only read one transcript.

  Ranked by

  Highest value: most agent surface types (MCP, CLI, API, SDK, plugin or skill, each counted once). A tie goes to the higher Features score, then to the higher Pricing score.

  Source

  The official pricing page, docs, and product pages.

  IDE / Coding Agent

  Qualifies

  The agent has a CLI or headless mode, and at least 4 extension points (MCP, skills, hooks, plugins, subagents, or an SDK).

  Ranked by

  Highest value: most ways to run the agent (desktop, CLI, IDE, web, mobile, chat, and code hosts). A tie goes to the higher Features score, then to more extension points.

  Source

  The official docs and product pages.

• Better for teams

  Qualifies

  A shared workspace with roles or post approvals, on a paid plan.

  Ranked by

  Highest value: lowest published price at which roles or approvals unlock. A custom price does not outrank a published one.

  Source

  The official pricing page.

• A category keeps at most 3 tools on a tag: the ones with the highest value on that tag's ranking. A tool that qualifies but ranks below them does not get the tag.
• A tool holds at most 3 tags, and can hold none. If it qualifies for more, it keeps the 3 where it ranks highest against the other tools.
• Every tag needs a source link and a one-line blurb of about 120 characters or less.
• A tag is re-checked when the source page changes or a new tool is added, and removed if the facts no longer hold.
• Tags never change the 0 to 10 scores. A score only breaks a tie. It never decides who qualifies.

Our pick for best value

On an alternatives page, the pick is the best step up from that tool. It is chosen for each tool, so different tools can get different picks.

1. Who competes. A tool with a free plan is compared only with other tools that have one. Any other tool is compared with paid tools priced up to 25% above its own entry price.
2. What matters most. Each of Pricing and Features counts more the further the tool’s own score is from 10. A tool weak on Features cares more about Features.
3. The gain. For each candidate, we take how much it beats the tool on each pillar, weighted as above, and average it. The highest gain wins.
4. Ties. Higher Pricing, then higher Features, then more shared switch reasons, then the earlier listed tool.

A tool with no score yet counts as 0 and never beats a scored one. If no candidate beats the tool, the page says so and shows the closest match. When one tool scores best on both pillars, it is the pick for every tool it beats.