Air: the address we list now redirects elsewhere · last checked 29 Oct 2026
Air
JetBrains Air is an agent development environment for running Claude Agent, Codex, Junie and Copilot against your real repos.
Air's bet is orchestration, not a model — and that is the right bet if you refuse to standardize on one vendor. Air Context plus event-triggered Automations solve the two things that actually hurt at scale: agents re-reading your repo and nobody knowing what each run cost. If you already pay for JetBrains AI Pro or Ultimate, access is included and the decision is close to free. If you don't, you are buying into the JetBrains frame — and cloud, web and mobile are still rolling out gradually, so
Verdict generatedLast checked 5d ago · cite: rightaichoice.com/tools/air
- JetBrains IDE users who want agents running inside the editor with in-context diff review
- Teams with mixed agent or model preferences that refuse to standardize on one vendor
- Platform and DevEx teams needing agent permissions, spend caps and usage analytics
- JetBrains AI Pro or Ultimate subscribers who get Air access with that subscription
- Solo developers on a single model who just want a terminal coding agent
- Teams standardized on VS Code or another non-JetBrains editor
- Anyone who needs every Air surface generally available today — cloud, browser and mobile rollout is gradual
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Air if you want a single opinionated coding agent rather than a control plane for several, or if your team is standardized on a non-JetBrains editor and won't benefit from the in-IDE half of the product.
Agent runs meter against a project AI credit quota — the site's own example shows a project at 300 of 1500 credits renewing Aug 1, so heavy automation use can exhaust the pool before renewal
Air's cost lands on a spectrum defined by what you already pay for. If you hold a JetBrains AI Pro or Ultimate subscription, agent access is included at no extra cost — the cheapest route by far. If you'd rather not, the BYOK path costs only what Anthropic, OpenAI or Google charge you for API usage, with no Air subscription. Compare against Cursor or GitHub Copilot, where you pay per seat for an editor plus bundled agents and don't get to swap model vendors mid-session.
In short
Air — JetBrains Air is an agent development environment for running Claude Agent, Codex, Junie and Copilot against your real repos. Best for JetBrains IDE users who want agents running inside the editor with in-context diff review, Teams with mixed agent or model preferences that refuse to standardize on one vendor, Platform and DevEx teams needing agent permissions, spend caps and usage analytics. Pricing not published.
No head-to-head comparison for Air yet — the alternatives below are the closest matches we list.
What is Air?
Air — JetBrains Air is an agent development environment for running Claude Agent, Codex, Junie and Copilot against your real repos. Best for JetBrains IDE users who want agents running inside the editor with in-context diff review, Teams with mixed agent or model preferences that refuse to standardize on one vendor, Platform and DevEx teams needing agent permissions, spend caps and usage analytics. Pricing not published.
Air is JetBrains' agent development environment — one system for building software with coding agents instead of a separate tool per model. It plugs into the agents you already pay for: Claude Agent, Codex, Junie, Copilot, OpenCode, and any agent you can wire up through ACP. You can connect compatible agents found on your machine with a click or pull new ones from an agent registry, and switch agents or subscriptions mid-session without losing your task. Where you work is your call. Install the plugin to run agents inside JetBrains IDEs, open air.jetbrains.cloud to work from any machine and share sessions with a team, or bring terminal agents like Claude Code and Codex into Air through Air Gateway. Cloud runs are live for some customers in IDEs and the browser today, with gradual rollout to everyone else over the coming months. Two pieces carry weight beyond a chat pane. Air Context semantically indexes your project so agents start with an understanding of the codebase rather than scanning it from scratch — JetBrains frames this as faster execution and lower cost on production-scale repositories. Automations fire agent runs on real events: merge to main, PR created, branch behind main, commit pushed, issue created. The site's own dashboard shows six running automations — an Agents.md updater at 1.18 credits, a code-reviewer at 18.46 — so cost per automation is visible, not theoretical. For teams, shared projects run in dedicated cloud environments with configurable VM size, internet access controls, and managed variables and secrets, plus MCP servers connected once for everyone's agents. Air Governance adds model and agent permissions by team or individual, flexible spend limits, bring-your-own Bedrock, OpenAI or Anthropic keys added once for all developers, and analytics on which tools teams actually use. It suits developers and orgs that have already decided agents are useful and now need several of them running without merge chaos.
Key features
- Run several coding agents in parallel sessions against real repositories
- Works with Claude Agent, Codex, Junie, Copilot, OpenCode and any ACP agent
- Switch agents or provider subscriptions mid-session
- Connect compatible agents found on your machine with one click
- Add agents from the built-in agent registry
- Air Context semantic project indexing for agent context
- Review agent diffs in context and leave line-level comments
- JetBrains IDE plugin for running agents inside the editor
- Browser access at air.jetbrains.cloud with shared sessions
- CLI via Air Gateway for terminal agents like Claude Code and Codex
- Automations triggered by PR created, merge to main, branch behind main, commit pushed, issue created
- Per-automation AI credit spend tracking
- Add files, folders, git branches, commits, local changes, symbols and terminals to task context
- Upload photos from phone into a task's context
- Shared team projects in dedicated cloud environments with VM size and internet access controls
Behind the verdict
Pick Air when the problem is not "which agent" but "too many agents." A team with Claude Agent for review, Codex for refactors and Junie for cheap bulk work currently stitches that together with terminal tabs and copy-paste context. Air collapses it into one session view where you can swap agents mid-session and review diffs in context, leaving line comments for the agent to act on. The feature that earns its keep is Air Context. Semantic indexing means a new session starts already knowing where components/Header.tsx lives and what depends on it, instead of burning tokens rediscovering the repo. JetBrains claims that lowers cost on production-scale codebases; even discounted heavily, fewer exploration tokens is a real saving on a large monorepo. Automations are the second reason to look. Wiring an agent to "PR created" or "merge to main" is ordinary CI work, but the dashboard makes each run's AI credit spend legible — the sample shows a code-reviewer at 18.46 credits and a lint checker at 4.52. If your CFO asks what agents cost, that number exists. Where it bites: the surfaces are uneven. Cloud runs are available to some customers in IDEs and the browser now, with everyone else waiting months. If your team lives in VS Code, this is the wrong door — the deepest integration assumes JetBrains IDEs, and the value thins out once you leave them. Solo developers on one model are over-served; Claude Code or Codex in a terminal costs nothing extra and skips the platform. The closest alternative is running agents directly, or Cursor and GitHub Copilot Workspace, which bundle an editor plus agents. Those ask you to standardize. Air's pitch is the opposite — keep your subscriptions and BYOK keys, add governance on top. The trade-off is the JetBrains frame: real org controls on
Models under the hood
as of 2026-10-24
Limitations
- Cloud runs are already available for some customers in JetBrains IDEs and the browser, with rollout to remaining customers happening gradually over the coming months, so you may not get the browser or shared-cloud experience on day one.
- Air Teams is early access and Air Governance is demo-gated.
- Agent usage is metered in AI credits against a project quota — the site's own example shows 300 of 1500 credits used before an Aug 1 renewal, and a flag counting members who exceeded quota.
- Work happens against real repositories in real environments, so review is expected rather than optional: you open diffs, validate with IDE tools, and comment on specific lines.
- The integrations page wasn't reached in this research pass, so no integration beyond those named on the main pages should be assumed or ruled out.
as of 2026-10-09
Researching Air? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Air still active?
Air: the address we list now redirects elsewhere · last checked 29 Oct 2026.
Viability score
How well maintained and how widely used is Air? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →What do users say about Air?
What people actually say about Air — is it worth it?
We scanned public community sources for Air on Aug 30, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Our own analysis of that scan says the posts were off-subject. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
How we verified this
Every listed fact about Air — pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 25 Oct 2026. Pages mix model-assisted synthesis with human review; see how we verify.
Verification history
We have re-verified Air 24 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed integrations, summary, features, our verdict, who it suits, description, pricing tiers, who should skip it
- — changed models
- — changed switching in, use cases, description, summary, our verdict, who should skip it, features, setup time, hidden costs, pricing fit, limitations, who it suits, pricing tiers, switching out, models
- — changed pricing fit, integrations, description, who should skip it, limitations, models, our verdict, who it suits, summary, switching in, features, setup time, use cases, switching out, hidden costs
- — changed SSL certificate
- — changed platforms, limitations, models, pricing page, website status
Showing the 6 most recent of 24 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Air cost in 2026?
Air: Pricing not published. Pricing last verified 25 Oct 2026.
Plans compared
Every published Air tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| JetBrains AI Pro / Ultimate | Included with subscription | — | Developers and teams already subscribed to JetBrains AI Pro or Ultimate who want agent access without a separate line itemAdds: Starting point for anyone with an existing JetBrains AI subscription — all supported agents included at no extra cost, including Claude Agent, Codex, Gemini CLI and Junie side by side. |
| Bring Your Own Key (BYOK) | Your API usage costs | — | Teams that already have Anthropic, OpenAI or Google contracts, or organizations centralizing keys from Amazon Bedrock for all developersAdds: No Air subscription required — you connect your own API keys or base URL and pay providers directly for usage, with JetBrains AI credits billed at public API rates as the alternative. |
| Air Teams | Early access | — | — |
| Air Governance | Custom | quote | — |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where Air's pricing actually pencils out — and where peers do it cheaper.
Air's cost lands on a spectrum defined by what you already pay for. If you hold a JetBrains AI Pro or Ultimate subscription, agent access is included at no extra cost — the cheapest route by far. If you'd rather not, the BYOK path costs only what Anthropic, OpenAI or Google charge you for API usage, with no Air subscription. Compare against Cursor or GitHub Copilot, where you pay per seat for an editor plus bundled agents and don't get to swap model vendors mid-session.
Who should pick Air?
Pick Air if you are JetBrains IDE users who want agents running inside the editor with in-context diff review, Teams with mixed agent or model preferences that refuse to standardize on one vendor, Platform and DevEx teams needing agent permissions, spend caps and usage analytics, JetBrains AI Pro or Ultimate subscribers who get Air access with that subscription.
Look elsewhere if you are Solo developers on a single model who just want a terminal coding agent, Teams standardized on VS Code or another non-JetBrains editor, Anyone who needs every Air surface generally available today — cloud, browser and mobile rollout is gradual, Teams that want a bundled editor plus agent in one product rather than an orchestration layer over agents they already pay for.
Skip it if Skip Air if you want a single opinionated coding agent rather than a control plane for several, or if your team is standardized on a non-JetBrains editor and won't benefit from the in-IDE half of the product.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Air actually fits — and what changes day-one when you adopt it.
Install the Air plugin, let Air find the agents already installed on the machine and connect them with a click, then run Claude Agent and Codex on the same task in parallel sessions.
Outcome: Both diffs open directly from their sessions, you validate changes with IDE tools, leave line comments on the version you prefer, and the agent revises — with each task showing its credit cost (0.58 credits and 0.31 credits in the site's own example).
Create a shared cloud project in Air Teams with a dedicated environment, invite the team, set VM size and internet access, add project secrets, and connect MCP servers once for everyone.
Outcome: Teammates run agent tasks in the same environment rather than on their own laptops, reach the same external tools through shared MCP connections, and access is managed from one place.
Stand up Air Governance, add an Amazon Bedrock key once so all developers get access, set model and agent permissions per team, and cap spend for teams and individual contributors.
Outcome: You can see which tools teams actually use and where AI contributes most, and you can redirect spend toward the workflows that pay off while flagging the members who exceed their quota.
Use cases
- Run the same task across Claude Agent, Codex and Junie and compare the diffs before merging
- Set up a code-reviewer automation that fires whenever a PR is created
- Delegate release notes, lint checks and issue triage to automations instead of doing them by hand
- Give a team shared cloud projects so they run agent tasks in the same dedicated environments
- Connect MCP servers once so every teammate's agents can reach the same external tools
- Bring an existing Claude or ChatGPT subscription or API key instead of buying new model access
- Track AI credit spend per automation to see which agent workflows are actually worth running
- Enforce model and agent permissions per team while capping AI spend at the org level
Setup time & first value
How long it actually takes to get something useful out of Air — broken out by persona, not the marketing-page minute.
In a JetBrains IDE: install the plugin and Air finds compatible agents already on your machine for one-click connection — first value typically arrives in the first working session. Browser access depends on cloud rollout, which is gradual, so timing varies by customer. Air Teams is early access and Air Governance is demo-gated, so org-wide setup should be scoped as a project with JetBrains
Switching to or from Air
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From terminal-only Claude Code or Codex: route the same CLI agents through Air Gateway so they run inside the same system alongside IDE and browser sessions
- →From a single-agent IDE plugin: install the Air plugin, let it detect the agents already installed, and keep working while adding Claude Agent, Codex, Junie, Copilot or OpenCode side by side
- →From per-developer API keys and shared credentials: add your Bedrock, OpenAI or Anthropic key once and let Air Governance distribute access with per-team permissions
- →From repo-by-repo prompt scaffolding: add files, branches, commits, local changes and MCP context to a task so agents start informed instead of scanning
- ↗To standalone Claude Code or Codex in the terminal: keep the same agents and drop the orchestration layer, losing parallel session tracking and shared cloud projects
- ↗To Cursor or GitHub Copilot Workspace: you get an editor plus bundled agents, but you lose mid-session agent swapping and the bring-your-own-provider model
- ↗To JetBrains AI Assistant alone: simpler in-IDE AI without Automations, shared cloud environments or org-level agent permissions
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “Air”, and we withheld 6: 6 could not be judged, because “Air” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Air.
Official links
What are the best Air alternatives?
The closest alternatives to Air we list are Cursor, emdash and Paseo. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Cursor | Freemium | Pricing not published | — |
| emdash | Free | Free | — |
| Paseo | Free | Free | — |
Alternatives to Air
View all
Tools that pair well with Air
Common stack mates teams adopt alongside Air, with the specific reason each pairing earns its keep.
Cursor
Cursor is an AI coding agent that plans, builds and ships features from your editor, CLI, Slack, GitHub and iOS.
emdash
Emdash is an open-source agentic development environment for running parallel CLI coding agents in isolated Git worktrees.
Paseo
Paseo is a self-hosted control plane for orchestrating Claude Code, Codex, Copilot and 30+ coding agents from desktop, phone or browser.
Best-of guides
Topics
Used Air? Help shape our editorial sentiment research.