JetBrains Air and Junie vs Claude Code and Codex in 2026: AI credits at $1, ACP agents and who pays for each run
JetBrains now prices AI in dollar credits, retires the Junie plugin on 3 December in favour of Air, and runs Claude Agent and Codex inside its IDEs. We read the licensing docs, the agent and ACP pages and the Junie blog to show who pays for each run, when your own Claude or ChatGPT plan applies, and where the brief and design states should live.

Your IDE now has a meter in dollars
If you write Kotlin, Java, Python or Go, there is a good chance you live in IntelliJ IDEA, PyCharm or GoLand. In 2026 JetBrains turned those IDEs into agent hosts: Junie, Claude Agent, Codex and GitHub Copilot sit in the AI chat, any Agent Client Protocol agent installs from a registry, and a new product, JetBrains Air, wraps it all into one system for building software with agents.
It also changed the unit. A JetBrains AI credit is now worth exactly one US dollar, charged at public API rates. AI Pro includes 10 of them every 30 days. That is honest pricing, but it is also one more meter, next to the Rovo credits in Jira and Confluence, the Figma AI credits, the Notion agent credits, the Miro pool and the Canva allowance.
That is the 2026 shape of AI credit fatigue: many small allowances, none of them shared, and context that hops between six tools before the agent writes a line. This guide compares JetBrains AI with running Claude Code and Codex on the plans you already pay, from JetBrains' docs and blogs checked on 22 November 2026, and shows where the shared context should live.
How JetBrains AI charges in November 2026
JetBrains AI has four tiers plus a trial. Prices differ for individuals and organizations, and every plan comes with a credit quota that resets every 30 days from the day you first used it.
| Plan | Price | Credits per 30 days | Notes |
|---|---|---|---|
| AI Free | $0 | 3 for individuals | Unlimited code completion and next-edit suggestions; no top-ups |
| AI Pro | $10 individual, $20 organization | 10 individual, 20 organization | Included in All Products Pack and dotUltimate |
| AI Ultimate | $30 individual, $60 organization | 35 individual, 70 organization | For heavy agent use |
| AI Enterprise | $60 organization | On par with Ultimate or higher | Organizations only; quota not disclosed |
| Top-up credits | $1 per credit | Valid 12 months | Trial, Pro and Ultimate; used after the included quota |
Three details matter. Credits are billed at public API rates, so a long agent run on a frontier model can use a Pro quota in an afternoon. Code completion and next-edit suggestions do not touch credits. And organizations on JetBrains Central share one pooled quota, with per-user limits set by an admin.
Junie is moving into Air
Junie is JetBrains' own coding agent. It left beta on 17 June 2026, and the company has been fast with it since: Gemini Flash defaults at a discount in August and September, Junie Local on your own machine on 24 August, and Junie Lite, a free hosted tier with very generous use, on 2 October.
- The plugin is retiring. The standalone Junie plugin is already off on 2026.3 IDEs and stops working on 2026.2 on 3 December 2026. Junie now lives as the coding agent in JetBrains Air, plus a CLI, a GitHub Action and GitLab CI/CD.
- Air is the new frame. Air runs Claude Agent, Codex, Junie, Copilot, OpenCode and other ACP agents, in the IDE, on the web, in a terminal and on mobile. Air in IDEs is in early access as a Marketplace plugin and in the 2026.3 EAP builds.
- Air without a JetBrains AI plan. JetBrains says you do not need a JetBrains AI subscription to use your own agents with Air. When an agent runs on a provider account, that provider's subscription, billing and terms apply.
- Junie BYOK at zero markup. With your own key Junie charges at the provider's rates, or you spend AI Pro and Ultimate credits. JetBrains claims up to 35 percent lower cost per task on the same model.
If your team installed Junie as a plugin last year, plan the move before 3 December. Our buyer's checklist for AI SDLC platforms lists the questions to ask any vendor whose agent product changed shape this year.
Claude Code and Codex inside JetBrains: three ways in
There are three different ways to run Anthropic's or OpenAI's agent in a JetBrains IDE, and they are billed differently.
- Claude Agent and Codex in AI chat. Built in, with sign-in through a JetBrains AI subscription, your own API key, a ChatGPT account for Codex or the Anthropic Console for Claude Agent. The method you choose is used for both processing and billing, so a ChatGPT sign-in or your own key does not touch JetBrains credits. Claude Agent reads CLAUDE.md and Codex reads AGENTS.md.
- Claude Pro or Max in Air's terminal view. A Claude subscription is not accepted in chat, because Anthropic's terms do not allow subscriptions through the SDK integration. In Air's terminal view it works, because Air runs the Claude Code you installed. This arrived in August 2026.
- ACP agents and the Claude Code plugin. Any agent from the ACP registry, or one you add to the acp.json file in your home folder, runs without a JetBrains AI subscription. Separately, Anthropic's own Claude Code [Beta] plugin, last released on 5 December 2025, connects the claude CLI to the IDE's diff viewer, selection and diagnostics. JetBrains notes it is not related to Claude Agent.
In practice: if you already pay for ChatGPT, sign Codex in with it. If you pay for Claude, run Claude Code in a terminal, in Air's terminal view or with the Anthropic plugin. Either way JetBrains credits stay untouched. The bring your own coding agent guide covers the three ways to pay for a run, and the Claude Code vs Codex comparison helps you pick the agent.
JetBrains AI, Claude Code and Codex side by side
All of these options can read a repository, plan, edit files, run commands and connect to MCP servers. They differ in who sells the model access.
| Option | Plans | How a run is paid | What it adds |
|---|---|---|---|
| Junie, Claude Agent or Codex on JetBrains AI | Pro $10, Ultimate $30 per individual | JetBrains credits at $1 each, API rates; top-ups valid 12 months | Agents in AI chat, IDE refactoring and inspections |
| Agents in JetBrains with your own key | Any JetBrains IDE; AI Free works | Your provider bills directly, zero markup | Same chat, no JetBrains credit |
| Claude Code in Air terminal or plugin | Claude Pro $20, Max $100, Ultra $200 | Your Claude plan | IDE diff viewer, selection and diagnostics |
| Codex in JetBrains or on its own | ChatGPT Plus $20, Pro $100 to $500, Business $25 | Your ChatGPT plan's Codex allowance | ChatGPT sign-in in AI chat, AGENTS.md |
| Sydla workspace | $19 per user | No AI credits: drives your own Claude Code or Codex, zero token markup | Brief, criteria, UX Studio states and evidence in one MCP connection |
Read the table by the plan you already pay. With a Claude or ChatGPT plan, JetBrains AI Free plus your own agent costs nothing extra; Pro is worth $10 mainly for JetBrains-hosted chat and Junie. If you own All Products Pack, AI Pro is already included.
The six-tool stack around the IDE
Whatever IDE hosts the agent, the run still needs a ticket, a brief, designed states, decisions and a launch visual. In a typical 2026 stack those live in six tools, each with a seat and an AI meter, and each an MCP hop away.
| Tool | Seat price | Bundled AI and its limit | Handoff to a coding agent |
|---|---|---|---|
| Jira Standard | $9.70 | Rovo: 25 credits per user per month, pooled; PAYG $0.01 per credit from 3 Dec 2026 | Atlassian MCP or a pasted ticket |
| Confluence Standard | $7.15 | Same Rovo pool, 25 credits per user (70 on Premium) | Atlassian MCP; enriched calls 1 to 10 credits |
| Notion Plus | $12.00 | Custom Agents on Business and Enterprise; extra credits $10 per 1,000 | Notion MCP server |
| Figma Professional Full | $20.00 | 3,000 AI credits per Full seat, no rollover | Figma MCP: 200 calls a day on Full or Dev |
| Miro Starter | $10.00 | 25 AI credits per license, pooled per team | Miro MCP: daily call limits by plan |
| Canva Business | $25.00 | Personal AI allowance per member | Outside the build loop |
| Sydla workspace | $19.00 | No AI credits: bring your own Claude Code or Codex | One MCP connection or a Sydla Desktop thread |
The six tools come to $83.85 per user per month before any overage, against $19 for Sydla, a saving of $64.85. Add JetBrains AI Pro for an individual and the stack is $93.85 per person with seven allowances to watch. The stack cost guide works through 5, 10 and 25 people, and the comparison overview covers each tool.
Technical walkthrough: IntelliJ, your agent and a shared brief
Sydla does not replace your IDE and does not add a meter. The context lives in one workspace, and the agent you already pay for reads it, whether you start it in IntelliJ, a terminal or Sydla Desktop.
- Brief. Each ticket sits in an initiative with a versioned brief, numbered acceptance criteria and a done-means line. Every save is signed by its author, a person or an agent.
- UX Studio. Screens and their states, empty, loading, error and success, are designed next to the ticket and published as a handoff with tokens and a URL per state.
- Wire MCP once. Register Sydla in the CLI: claude mcp add with the HTTP transport and the sydla.com/mcp URL for Claude Code, or an mcp_servers.sydla table in config.toml for Codex. In the IDE, AI Assistant's MCP settings accept the same server over Streamable HTTP, and can pass your MCP servers to ACP agents. The IDE's own MCP server, built in since 2025.2, then lets Claude Code or Codex use IntelliJ's inspections and refactorings too.
- Run in Sydla Desktop. Open the ticket, start a thread and choose claude or codex. Desktop drives the CLI you are already signed in to, so the model, limits and bill stay on your plan. Each thread has an access level, Read only, Edit files or Full access, with a Preview panel and a Changes view. Keep IntelliJ open on the same worktree to review the diff with its inspections.
- Review. The agent records evidence against each criterion: test output, screenshots of each state and the diff. A person reviews it and decides the merge and release gates. An agent can request a gate but never approve one, and the record stays with the ticket rather than in an IDE chat.
The brief to pull request walkthrough shows one feature end to end, and the Sydla Desktop setup guide covers the install and sign-in for both CLIs.
Five checks before you standardise on JetBrains AI
JetBrains IDEs are still the best place to read and refactor a large JVM or Python codebase. Before you buy AI seats for the team, check five things.
- Which plan pays. If everyone has a Claude or ChatGPT plan, AI Free plus your own agent costs nothing extra. Pro buys 10 dollars of hosted use; check whether All Products Pack already gives you Pro.
- Credit burn. One credit is one dollar at API rates. Watch the quota on agent-heavy days; top-ups last 12 months, but the included quota resets every 30 days.
- Claude plan terms. A Claude subscription works in Air's terminal view and with the Claude Code plugin, not in AI chat. Claude Agent in chat needs JetBrains AI, an Anthropic API key or the Anthropic Console.
- The Junie move. The Junie plugin stops on 2026.2 on 3 December. Move to Air, the CLI or the CI integrations before then.
- Where context lives. An IDE does not fix a brief spread across five tools. Decide where the ticket, criteria and states live before you decide where the agent runs.
When JetBrains AI is the right choice
Be fair to JetBrains. It is the right choice if your team already pays for All Products Pack, wants agents that can call real IDE refactorings and inspections, and likes having Junie, Claude Agent, Codex and Copilot behind one chat with a shared organization pool and admin limits. Junie Lite and Junie Local make the free path stronger than it was a year ago.
It is a weaker fit if your team mostly runs agents from a terminal, or if the real problem is not the IDE but a brief that lives in five places. The Zed comparison, the Cursor and X Subscription pricing guide and the AI credit fatigue guide look at the same question from other sides.
Try it on your next ticket
Sydla is $19 per user per month with every feature included and no AI credits to buy. Your Claude Code or Codex plan does the coding, in IntelliJ, a terminal or Sydla Desktop, and the brief, the states and the evidence live with the ticket. See the pricing section, read how Sydla works, then start your workspace and keep your Claude Code and Codex accounts.
Frequently asked questions
How much does JetBrains AI cost in 2026?
For individuals, AI Free is $0 with 3 credits, AI Pro is $10 with 10 credits and AI Ultimate is $30 with 35 credits per 30 days. For organizations, Pro is $20 and Ultimate and Enterprise are $60. One credit equals $1 at API rates, as of November 2026.
Can I use Claude Code or Codex in IntelliJ IDEA?
Yes. Claude Agent and Codex are built into AI Assistant chat, Codex can sign in with a ChatGPT account, and any ACP agent installs from the registry. A Claude Pro or Max plan works through Air's terminal view or Anthropic's Claude Code plugin, not in AI chat.
Is Junie being discontinued?
The Junie plugin is. It is off on 2026.3 IDEs and stops working on 2026.2 on 3 December 2026. Junie itself continues as the coding agent in JetBrains Air and as a CLI, a GitHub Action and a GitLab CI/CD integration.
Do Claude Agent or Codex use JetBrains AI credits?
Only if you sign them in with your JetBrains AI subscription. With your own API key, a ChatGPT account or the Anthropic Console, the provider handles both processing and billing, and ACP agents need no JetBrains AI subscription at all.
Do I still need Jira, Figma and Notion if my agent runs in a JetBrains IDE?
The agent still needs a ticket, a brief and designed states. In a six-tool stack that is $83.85 per user a month with six AI meters. Sydla holds the brief, criteria, UX Studio states and evidence for $19 per user and runs your own Claude Code or Codex with zero token markup.



