Linear AI Review Australia 2026: Skip Unless You Build Software

Verdict upfront. Linear AI is genuinely good. It's also almost certainly the wrong tool for the person reading this review. Linear is an issue tracker for engineering and product teams, and Linear AI is built around that job. If you run a marketing agency, accounting practice, property business or any team that doesn't ship software, skip this one and go straight to ChatGPT, Claude or Notion AI. The rest of this review explains why we still gave it three stars, and who should actually buy it.

What is Linear AI?

Linear is the issue tracker used by OpenAI, Coinbase, Ramp and a long list of well-known software companies. Sprints, backlogs, bug reports, GitHub pull requests, code reviews. Linear AI is the AI layer the company has been building on top of that since mid-2025. It includes Linear Agent (chat with workspace context, available in-app, in any comment via @Linear, and inside Slack), Triage Intelligence (auto-routes incoming bug reports and feature requests), Linear Asks (turns Slack messages into proper tickets), and Code Intelligence (lets non-engineers ask questions of your codebase without interrupting a developer). There's also a Linear Agent Interaction SDK that lets third-party agents like Cursor, OpenAI Codex and Devin live inside a Linear workspace as first-class users with their own permissions.

The vendor is Linear Orbit, Inc. — a Delaware corporation, San Francisco-headquartered, founded in 2019 by former Coinbase and Airbnb design and engineering leads. Backed by Sequoia and Accel.

The product is excellent at what it does. The catch is the tool it lives in. If Linear isn't already your team's home, none of this AI matters to you.

What we like

Linear names its model stack — most peers don't. The place it does so is the Coding sessions docs page, and it is specific: sessions run “through Claude Code or Codex”, the default model tracked under Auto is Claude Opus 4.8, and a workspace admin can switch to Claude Fable 5, Claude Opus 5, Claude Sonnet 5, GPT-5.6 Sol, GPT-5.5 or GPT-5.4 (checked 21 September 2026). That is unusual transparency in a category that mostly hides behind “multi-model routing”, and it lets you make an informed call about whether the underlying providers fit your privacy posture.

But the transparency does not extend to the other two AI features, and an earlier version of this review said it did. Linear's Code Intelligence docs page and its 14 May 2026 changelog entry name no model at all — the Claude Code attribution belongs to Coding sessions, which is a different feature. Triage Intelligence's docs page says only that issues are “analyzed by agentic models”. The GPT-5 and Gemini 2.5 Pro pairing we used to quote comes from a Linear engineering post dated 3 September 2025, which describes the models Linear had just switched to and says plainly that Triage Intelligence will improve “as we adopt newer models and techniques”. Treat it as build history, not as a statement of what runs today.

Code Intelligence is the strongest feature in the suite. A product manager, designer or customer-support lead can ask questions about the codebase without filing a Slack ticket to engineering. "Where does this user's billing data get written?" gets a real answer, with file references. Unique among workspace-AI peers — Notion AI, ClickUp Brain, Rovo and the others have nothing equivalent. Free during public beta on Business and Enterprise plans.

Linear Agent is on the Free tier. Most workspace-AI vendors gate their flagship AI feature behind a paid plan. Linear includes Linear Agent (beta) on every plan including Free, with "an amount of Agent usage designed to cover typical day-to-day work." Generous for a team kicking the tyres.

The Agent Interaction SDK is a real platform play. Third-party coding agents — Cursor, OpenAI Codex, GitHub Copilot, Devin, Sentry's Seer, Warp, Factory, Charlie — install as first-class workspace users with their own OAuth scopes and audit trail. You can hand a ticket to Cursor the way you'd hand it to a human, and Cursor opens a pull request against the linked issue. Nobody else in the category has built this layer.

Training-data policy is clean. The Triage Intelligence FAQ states directly: "Linear does not utilize your data to train its own AI models. Any data processed to enable Linear's AI features is shared with our trusted partners… exclusively to deliver those AI functionalities to you without permission to train on provided data."

Watch out for

A real reliability incident in March 2026. On 24 March a code change meant to skip permission filters in one internal test workspace shadowed a variable and caused every workspace on the platform to lose team-level permission filtering for about an hour (12:07am to 1:10am UTC). Workspaces with private teams plus guest accounts were the most exposed. Affected workspaces were notified by email on 26 March; Linear published a detailed post-mortem on 6 April. The disclosure was thorough — and they note pointedly that the buggy code was written and reviewed by engineers without AI assistance. Still, exactly the sort of incident a regulated AU business should weigh before sending sensitive material into a US-hosted workspace.

USD-only pricing with no published Australia-specific GST guidance. Linear's pricing page is USD across the board. There's no AUD toggle. Stripe handles checkout — Stripe Tax supports AU GST collection and ABN-exempt reverse-charge when enabled by the merchant, but Linear itself doesn't document whether they have it switched on. The underlying mechanic is standard global-SaaS, but verify on your first invoice rather than trust an assumption — and if GST handling matters to your bookkeeper, ask Linear support directly before scaling seats.

No Australian data residency. Workspace creation offers two regions: United States or European Union. There is no Sydney option, and no public roadmap for one. The same applies to the AI sub-processors listed in Linear's DPA — Anthropic, OpenAI, Fireworks, Cohere, AWS, Google — all of which sit in US or EU regions. For regulated workloads, choose EU at workspace creation.

No IRAP, no Hosting Certification Framework, no Essential Eight self-attestation. Linear holds SOC 2 Type II, ISO/IEC 27001:2022, GDPR and HIPAA (BAA available on Enterprise). All useful. None of them are the Australian Government's certification of choice. If you sell to Australian Government or work in a sector that defaults to IRAP, Linear isn't on the panel.

Usage-based pricing has now arrived. The Linear Agent docs always flagged it — "Some capabilities that require more compute, like Automations, may move to usage-based pricing when they reach general availability" — and as of 11 June 2026 the first piece has landed. Coding Sessions (Linear Agent writing code with Claude Code and Codex) runs on AI credits on Basic, Business and Enterprise; workspace admins manage usage and credits in Settings. Source Code Intelligence is still free during its beta on Business and Enterprise, and will be priced at general availability. Source The seat price hasn't moved, but credit consumption is now a real variable for anyone planning to standardise on Linear's agentic features.

Audit log retention is only three months. Linear's security page states audit logs cover "the last 3 months." That's short for many AU compliance regimes. Worth knowing if you ever need to reconstruct activity for a regulator or insurer.

Pricing

Linear's pricing page shows USD only, billed annually. Monthly billing is selectable in-app at a higher rate that the public pricing page does not display. AUD figures below are calculated at a 1.40 USD-to-AUD rate as of 23 May 2026 and treated as estimates.

Value verdict: Business at ~AUD$27/seat/month (estimated, inc. GST) is fair against the workspace-AI peer set — Notion AI Business sits at ~AUD$27, ClickUp Brain at ~AUD$15 on top of Unlimited. For an engineering team already on Linear, Business is good value. For everyone else, the question isn't "is it worth $27" — it's "should you be buying workspace AI at all, or just ChatGPT or Claude." For the wrong-fit reader, the answer is the second one. Linear publishes no Australia-specific GST guidance — Stripe handles checkout, so ABN-exempt should work, but check your first invoice rather than trust an assumption.

Australian notes

This is where Linear scores lowest in the workspace-AI category.

• Available in Australia: Yes. No geographic restriction on signup.
• Pricing in AUD: No. USD only via Stripe, FX applied by your bank.
• GST: Not addressed on Linear's pricing or billing pages. Linear uses Stripe for checkout; Stripe Tax supports AU GST collection + ABN-exempt reverse-charge when the merchant enables it, but Linear doesn't confirm whether they have. Standard global-SaaS pattern is likely — verify on your first AU invoice.
• Data residency: None local. United States or European Union only, chosen at workspace creation. No Sydney. No published roadmap.
• AI inference residency: None local. Every AI sub-processor sits in US or EU regions per the DPA.
• Compliance: SOC 2 Type II, ISO/IEC 27001:2022, GDPR, HIPAA. No IRAP. No Hosting Certification Framework listing. No Essential Eight attestation.
• Local support: None. Priority support is Enterprise-only. No AU phone, no AU office, no AU-hours chat.
• AU-aware outputs: None. Linear Agent defaults to US English. No AU templates. No AU-business context in the product.
• Forward-looking signal: Linear has an open Account Executive APAC role based in Australia on its careers page. First sales hire on the ground here. Encouraging — but it's a job ad, not a delivered product.

Australian Fit score: 1 out of 5. The lowest AU fit in the workspace-AI peer set we've reviewed in this cohort.

Best for

• Australian engineering teams already on Linear who want best-in-class issue-tracker-native AI. Code Intelligence alone is worth the Business upgrade for a 10+ person team.
• Product teams orchestrating work between humans and third-party coding agents (Cursor, Codex, Devin). The Agent Interaction SDK is genuinely unique.
• Software companies running on Slack that want intake from any channel — Linear Asks Agent now files a structured ticket from a @Linear Asks mention in any Slack channel.
• Engineering leaders willing to host in the EU region to bring AI inference closer to a defensible compliance posture.

Not great for

• Non-developer small business owners. If you don't ship code, Linear isn't your workspace, and Linear AI doesn't help you do your job.
• Finance, accounting and bookkeeping teams. Buy Notion AI, ClickUp Brain or stay in your accounting platform.
• Marketing teams and creative agencies. A general assistant (ChatGPT, Claude) plus a content tool (Jasper, Copy.ai) does more for less.
• Sole operators. The per-seat economics make sense for a team. They don't make sense for one person who just wants AI help with their inbox and documents.
• Australian Government suppliers or any IRAP-required workload. Not on the list, no roadmap.
• Any business that requires Australian data residency. US and EU only, full stop.

Alternatives worth considering

For most readers of this review, the honest answer is the first row. If you don't already run on Linear, leave the workspace-AI category and buy ChatGPT or Claude. Faster to set up, cheaper, and does more of what you actually need.

Our verdict

Linear AI is genuinely good — for engineering and product teams. The model stack is named, Code Intelligence is unique in the category, Linear Agent is on the Free tier and the Agent Interaction SDK is the most serious platform play in workspace AI right now. If Linear is already your team's home and you're on Business or Enterprise, it earns its keep.

For everyone else, this is the wrong category. Buy ChatGPT or Claude, save the seat fees and skip Linear AI entirely. A Melbourne business owner running a six-person agency, a property practice or an accounting firm is not the buyer here. The right tool for that reader is a general assistant, not an engineering issue tracker with an AI layer bolted on.

If you're an Australian engineering leader weighing Linear: choose the EU region at workspace creation, verify the GST mechanic on your first invoice (Stripe-backed checkout, but vendor-unconfirmed for Australia), factor in the March 2026 permission-filter incident when you decide what data goes into private teams, and assume Code Intelligence and Agent automations will move to usage-based pricing at GA.

Scores

Reviewed by The AI Directory Australia. Last updated 28 June 2026 (sourced from verified dossier dossier-linear-ai-2026.md). Pricing and features verified against linear.app/pricing. USD-AUD estimates at 1.40. Subject to change — verify on the vendor page before purchasing.

Frequently asked questions

Is Linear AI available in Australia?

Yes, with no geographic restriction on signup. But it's the wrong tool unless you build software — if you don't ship code, buy ChatGPT or Claude instead.

Does Linear charge in AUD?

No. Linear's pricing page is USD only via Stripe, with FX applied by your bank. AUD figures are estimates at a 1.40 rate. Business runs roughly A$27/seat/month inc. GST.

Is there a free plan or trial?

Yes. Linear has a Free tier that includes Linear Agent (beta) and MCP access on every plan. The flagship AI features (Code Intelligence, Triage Intelligence) need the Business plan.

Does Linear store data in Australia?

No. Workspace data is hosted in the United States or European Union only, chosen at creation — no Sydney region and no roadmap for one. For regulated workloads, choose EU. There is no IRAP listing.

Who is Linear AI best for?

Australian engineering and product teams already on Linear. Code Intelligence and the Agent Interaction SDK are best-in-class. Non-developers, finance teams, agencies and sole operators should skip it.

Read next

• Category guide: Best Workspace AI for Australian Teams 2026
• Reclaim AI review
• Taskade review

Sources

1. Linear / Pricing (Primary — pricing, plan tiers, AI feature placement, USD-only, Enterprise-only support & SLA; Free $0 / Basic $10 / Business $16 confirmed unchanged) — https://linear.app/pricing — checked 28 June 2026
2. Linear / Security (Primary — data residency regions, SOC 2 / ISO 27001 / GDPR / HIPAA, multi-region hosting, 3-month audit log retention) — https://linear.app/security — checked 13 June 2026
3. Linear / Data Processing Addendum (Primary — AI sub-processor list, US/EU regions, Stripe payments processor) — https://linear.app/dpa — checked 13 June 2026
4. Linear / Linear Agent docs (Primary — Linear Agent availability, usage-based pricing flag for automations, MCP access, agent skills) — https://linear.app/docs/linear-agent — checked 13 June 2026
5. Linear / Triage Intelligence docs (Primary — Triage Intelligence feature, training-data policy quote) — https://linear.app/docs/triage-intelligence — checked 13 June 2026; re-read 21 September 2026: the page no longer names any model, only “agentic models”. Nearest existing Wayback capture 3 September 2026.
6. Linear / Code Intelligence docs (Primary — Code Intelligence still public beta + free during beta, Business/Enterprise only) — https://linear.app/docs/code-intelligence — checked 28 June 2026; re-read 21 September 2026: the page names no underlying model. Nearest existing Wayback capture 12 June 2026. 6a. Linear / Coding sessions docs (Primary — “we start a secure coding session through Claude Code or Codex”; Auto default Claude Opus 4.8; selectable Claude Fable 5, Claude Opus 5, Claude Opus 4.8, Claude Sonnet 5, GPT-5.6 Sol, GPT-5.5, GPT-5.4; AI-credit usage) — https://linear.app/docs/coding-sessions — checked 21 September 2026 — nearest existing Wayback capture 19 September 2026, two days before our read; archive.org is still refusing new saves.
7. Linear / Linear Asks (Primary — Linear Asks feature, Slack intake) — https://linear.app/asks — checked 13 June 2026
8. Linear / Linear for Agents (Primary — Agent Interaction SDK, third-party agent list: Cursor, Codex, Devin, Copilot, etc.) — https://linear.app/agents — checked 13 June 2026
9. Linear / Changelog (Primary — 21 May 2026 Linear Asks Agent; 14 May 2026 Code Intelligence) — https://linear.app/changelog — checked 13 June 2026
10. Linear / "How we built Triage Intelligence" engineering post, published 3 September 2025 (Primary, but historical — describes the v1 GPT-4o mini + Gemini 2.0 Flash pairing and the switch to GPT-5 + Gemini 2.5 Pro as they stood then, and states Linear will “adopt newer models and techniques”) — https://linear.app/now/how-we-built-triage-intelligence — checked 13 June 2026; re-read 21 September 2026 — Wayback 31 August 2026
11. Linear / Incident post-mortem, 24 March 2026 (Primary — permission-filter incident, variable shadowing, ~1 hour outage, 26 March email notification, 6 April post-mortem) — https://linear.app/now/linear-incident-on-mar-24th-2026 — checked 13 June 2026
12. Linear / Careers (Primary — open Account Executive APAC role based in Australia) — https://linear.app/careers — checked 13 June 2026
13. Linear / Changelog — Coding Sessions, 11 June 2026 (Primary — Coding Sessions on Basic/Business/Enterprise, AI credits / usage-based, Claude Code + Codex) — https://linear.app/changelog/2026-06-11-coding-sessions — checked 28 June 2026
14. Linear / Changelog — Linear Diffs, 28 May 2026 (Primary — native code review on all plans, Guided Reviews beta on Business/Enterprise) — https://linear.app/changelog — checked 28 June 2026