Tools, and what they cost this month.
Prices and limits change every few weeks, so every row says when we read it and links to the page we read. USD per month before tax, as the official pages show them. We earn nothing from any of these links.
If you've never coded: start in an app builder (Lovable, Replit, Bolt, v0) for a weekend, then move to an agent that works on files you own (Claude Code, Codex, Cursor) when you want to keep the code. If you code already: the plan matters less than how you use it. The tickets for Junior Dev and Tech Lead are about that.
| Tool | Good for | Free | Cheapest paid | Other plans | How usage is limited | Checked |
|---|---|---|---|---|---|---|
| Claude CodeAnthropic · Agent in the terminal | Long, multi-file work from the terminal where you want the agent to read the repo, run commands and fix its own mistakes. Less good for: People who want a visual editor or a hosted app builder; you still need to know where your code lives and how to run it. | No free access to Claude Code. The Free plan covers chat only. | $20 Pro ($17/month billed yearly ($200 up front)) |
| Usage is capped per 5-hour session and per week, as multiples of Pro. Ultra drops the 5-hour cap for verified accounts but keeps the weekly one. Limits also depend on model, chat length and file size. | 19 Nov 2026 source |
| OpenAI CodexOpenAI · Agent in the terminal | Developers already paying for ChatGPT who want an agent in the terminal, the IDE and the cloud under one subscription. Less good for: Predicting how far a plan goes: limits are ranges per model, not fixed numbers. | Codex is included in ChatGPT Free, with limits the page does not state. | $8 Go |
| Plus and standard Business get a range of local messages per 5-hour window for each model (GPT-6 Luna 350 to 3,000, GPT-6.2 Astra 6 to 50). The page says these are not fixed limits and weekly limits may also apply. Recent change: Ultrafast overages on GPT-6.2 Astra and Sol cost 3x the standard rate through 2027-02-28 and 6x from 2027-03-01, per the pricing page. | 19 Nov 2026 source |
| CursorAnysphere · Editor with an agent | People who want an AI editor and are happy to use Grok models, which are the only ones on the cheapest plan. Less good for: Anyone who mainly wants Claude or GPT: on Premium and up those are billed separately at API list price. | Free plan with limited agent and chat requests and limited Grok usage. | $8 Plus (Grok models only) |
| One usage pool shared across Cursor and Grok products, refreshed weekly. You can buy top-offs or wait for the refresh. Third-party models are billed at API list price and do not come out of the pool. | 19 Nov 2026 source |
| Devin DesktopCognition · Editor with an agent | Existing Windsurf users and teams who want a local editor plus Devin cloud agents on one bill. Less good for: Budgeting: the page gives no numbers for any plan's quota. | Light agent quota with limited models, plus unlimited tab completions and inline edits. | $20 Pro |
| Paid plans get an allowance that refreshes daily and weekly. Extra usage is bought at API pricing. No quota figures are published. Recent change: Windsurf became Devin Desktop on 2026-06-02 as an over-the-air update, with pricing unchanged (Devin docs FAQ). | 19 Nov 2026 source |
| GitHub CopilotGitHub (Microsoft) · Editor with an agent | Cheap entry point if you already live in VS Code or GitHub; $10 gets agent mode and the cloud coding agent. Less good for: Heavy agent use on Pro: $15 of credits runs out quickly once agent mode and code review start billing. | 2,000 code completions per month plus limited chat and agent usage. Agent mode included; the cloud coding agent is not. | $10 Pro (per user) |
| Completions are free on paid plans. Agent mode, code review, the cloud agent, the CLI and the Copilot app spend AI credits (1 credit = $0.01). Extra credits can be bought on paid plans with a spending budget. Recent change: All Copilot plans moved to usage-based billing with AI Credits on 2026-06-01, replacing premium requests (GitHub blog, 2026-04-27). | 19 Nov 2026 source |
| Google AntigravityGoogle · Editor with an agent | Trying an agent-first editor for free, and the Google AI Pro plan includes Claude Sonnet and Opus 5.5 alongside Gemini. Less good for: Knowing your limits in advance: no plan states a numeric quota. | $0 with a baseline agent quota refreshed weekly, on Gemini 3.8 Flash (older Claude 4.6 and GPT-OSS models were listed for removal on 2026-11-18). | $19.99 Google AI Pro ($199.99/year billed annually (US price on Google One)) |
| Free: weekly quota. AI Pro and Ultra: quota refreshed every 5 hours plus weekly caps. None of the tiers publishes numbers. Recent change: Claude Sonnet 4.6, Claude Opus 4.6 and GPT-OSS-120b were listed for removal from Antigravity on 2026-11-18, per the pricing page. | 19 Nov 2026 source |
| LovableLovable · App builder in the browser | Non-programmers who want a working web app with a database and hosting from a written description. Less good for: Big or unusual codebases; 100 credits a month goes fast when the app gets complicated. | 5 build credits a day (up to 30 a month), 20 Cloud credits a month, publishing limited to invite-only. | $25 Pro (yearly billing advertised as '2 months free'; yearly price not shown) |
| Credit-based: each build request spends credits. Daily grant plus a monthly allowance; Cloud credits are separate. | 19 Nov 2026 source |
| ReplitReplit · App builder in the browser | Building and hosting an app entirely in the browser, with the agent, database and deploys in one place. Less good for: Cost control on bigger projects: Pro scales up to $2,000 a month and usage is billed by effort. | The pricing page does not list a free plan. | $20 Core ($18/month billed annually) |
| Monthly credits for paid models, plus a 'Free Mode' with hour and project caps. Effort-based pay-as-you-go after credits run out. | 19 Nov 2026 source |
| Bolt.newStackBlitz · App builder in the browser | Quick front-end prototypes in the browser with a generous free token allowance. Less good for: Large apps: token use grows with project size, and every request re-reads more code. | 1M tokens a month with a 300K daily cap, Bolt branding on sites. | $25 Pro (billed monthly; yearly billing 'saves up to 28%') |
| Token-based. Free tokens do not roll over; Pro tokens do. | 19 Nov 2026 source |
| v0Vercel · App builder in the browser | React and Next.js interfaces that you plan to deploy on Vercel anyway. Less good for: Back-end heavy apps, and anyone on the free plan who needs more than 7 messages a day. | 7 messages a day, deploys to Vercel, GitHub sync. | $30 Plus (per user) |
| Credit-based, with token rates per model. Extra credits can be bought and shared across a team. | 19 Nov 2026 source |
Names belong to their owners. Lock In Lab isn't affiliated with any of these companies. Something out of date? The source link in each row is the page to compare against.