AgentGid

Live data on 250+ AI agents: rankings, prices, changelogs, outages and 40K MCP servers. Which agent to use and what changed this week.

[AgentGid @agentgid]

Choosing an AI agent shouldn't take a week of tabs.

AgentGid tracks 250+ of them with live public data:

- find yours in 3 questions
- compare any two side by side
- see if it's down right now
- what changed this month
- 40K MCP servers, ranked

Free: https://t.co/LkSWi2ltr9

[AgentGid @agentgid]

Merged pull requests from AI coding agents on public GitHub last week: Claude Code leads with 629K. https://t.co/Sr4qAZJlQw

@claudeai @OpenAIDevs @cursor_ai

[AgentGid @agentgid]

Databases, ORMs and clients in projects built with AI coding agents: pg is in 5% of 6,211 public repos, Drizzle ORM in 4%. Full ranking: https://t.co/V09o2MY3LM

[AgentGid @agentgid]

New AI agents worth a look: Rea added 42K GitHub stars in 7 days, Orca 6,085. Live movers: https://t.co/mVxQG95Cn4

@orca_build @deepseek_ai

[AgentGid @agentgid]

Which MCP servers do developers actually configure? We read the agent configs of 10,078 public repos: Context7 is in 15% of them, Playwright MCP in 14%. https://t.co/0zU7De8rqf

[AgentGid @agentgid]

10 AI agents changed their prices this week: OpenHands, https://t.co/UVrlups1MX, Kortix, https://t.co/IzZlLOlIZP…. All prices side by side: https://t.co/MnLixYmobz

@OpenHandsDev @boltdotnew @kortix

[AgentGid @agentgid]

One hacker's mostly autonomous AI agents breached 100+ websites, Norforge finds

Broke into more than 100 websites with little involvement from the human hacker, per Norforge
Hit companies in the US, Canada, Australia and Japan, per Reuters
https://t.co/75EeXnFGfj

[AgentGid @agentgid]

How often did AI coding agents report problems in the last 30 days? Cursor posted 18 incidents on its status page. Month-by-month reliability: https://t.co/ipJgkuH4Ee

@cursor_ai @Lovable

[AgentGid @agentgid]

Cursor drops its entry plan from $20 to $8 a month, a 60% cut

The entry price went from $20 to $8 per month, a 60% reduction. For one seat that is $144 less per year. The facts we have do not explain why Cursor changed the price.

[AgentGid @agentgid]

Merged pull requests from AI coding agents on public GitHub last week: Claude Code leads with 582K. https://t.co/Sr4qAZIO0Y

@claudeai @OpenAIDevs @cursor_ai

[AgentGid @agentgid]

Read more: https://t.co/jEa1HNuZix

[AgentGid @agentgid]

AI Content Factory: 10 Open-Source GitHub Repos, Checked
A viral post lists 10 open-source repos for an AI content factory. We checked each: code and weights licenses, activity, the paid APIs several depend on, and which ones download from social platforms.

[AgentGid @agentgid]

Testing in projects built with AI coding agents: Vitest is in 20% of 6,152 public repos, Playwright (JavaScript) in 17%. Full ranking: https://t.co/f15p0kdD9r

[AgentGid @agentgid]

How often did AI coding agents report problems in the last 30 days? Cursor posted 19 incidents on its status page. Month-by-month reliability: https://t.co/ipJgkuH4Ee

@cursor_ai @Lovable

[AgentGid @agentgid]

Cursor pricing changed: Free + $20/mo → Free + $8/mo. Plans and limits: https://t.co/Ehyqki2zII

@cursor_ai

[AgentGid @agentgid]

@skunzu We broke this down against Anthropic's docs and prices: the advisor is real, but each call re-reads the full transcript at Opus price, so it only saves money when the context stays small. Numbers here: https://t.co/vpzAzBA2Ey

[AgentGid @agentgid]

Sonnet builds. Opus reviews. Haiku reads.

Claude Code's advisor setup (h/t @skunzu) is real, but not free:

→ each advisor call re-reads the whole conversation at full Opus price
→ advisor alone: $4.90 vs $4.73 for plain Opus
→ with Haiku helpers keeping context small: $2.82

[AgentGid @agentgid]

@skunzu What matters more than the model mix: hand the agent the plan, the known traps and the checks before it starts. That is exactly what the advisor is paid for.

Full breakdown, Anthropic's prices and a fact-check of the original post:
https://t.co/bkgKtBTtS1

[AgentGid @agentgid]

Sonnet builds. Opus reviews. Haiku reads.

Claude Code's advisor setup (h/t @skunzu) is real, but not free:

→ each advisor call re-reads the whole conversation at full Opus price
→ advisor alone: $4.90 vs $4.73 for plain Opus
→ with Haiku helpers keeping context small: $2.82

[AgentGid @agentgid]

UI components and icons in projects built with AI coding agents: Lucide React is in 19% of 14,835 public repos, Class Variance Authority in 14%. Full ranking: https://t.co/kyj7rrv55G

[AgentGid @agentgid]

@iamvishal16_ios @OnlineGid @ClaudeDevs Real resets only. We count one when the Claude Code or Codex team officially announces it on X, not plan-page dates. You can also turn on alerts and get notified the moment a new reset drops: https://t.co/5ZTdLFq6fs

[AgentGid @agentgid]

@ClaudeDevs it's been 18 days since the last Claude Code limit reset, and Codex users got 6 resets in the last 30 days. Could we get one too? 🙏 https://t.co/LGih0hunxT