Subscribe

Matrix/Method

Method and sources

AI Coding Agents Tracker compares 16 coding agents and app builders on 11 features. 116 of 176 cells are verified from 40 vendor pages and the official SWE-bench data file. Last checked Nov 4, 2026.

How a cell is filled

What Unknown means

We could not confirm the value on a page we fetched. That includes pages that render prices only in JavaScript (Lovable's plan prices, for example). Unknown is never a No.

Scope

Cloud coding agents work on their own on a repo and hand back a diff or PR. Prompt-to-app builders turn a description into a hosted app. IDE completion and in-editor agent mode are covered on our sister site, the AI Integration Matrix.

How this differs from other comparisons

This tracker puts cloud agents and app builders in one filterable matrix, sources every cell, adds official SWE-bench entries and publishes a diff-based update feed and JSON.

Refresh

Each week: refetch sources, rebuild data/agents.json, run scripts/gen_updates.py to diff against the last snapshot, then rebuild and deploy.

Corrections

Spotted something out of date? Changes are made only from a public source. Send the source link via the newsletter reply address once you've subscribed.

Affiliate and sponsor disclosure

Some outbound links may become affiliate links, and one weekly sponsor slot is sold (see Advertise). Neither changes any cell: values come only from sources.

Weekly, free

Which coding agent changed this week

One short email when coding agents and app builders change prices, limits, MCP support or benchmark results. Every item links its source.

You'll finish signing up on beehiiv.