# Linear AI automation and integrations > Linear has a GraphQL API with personal API keys or OAuth 2.0, signed webhooks, and built-in integrations in n8n, Zapier and Make. Good AI automations triage bugs from support, turn customer feedback into linked issues, summarize alerts and write project updates. Linear also supports AI agents as app users that can be assigned issues. API keys allow 2,500 requests per hour per user. - Category: Project Management - Vendor: https://linear.app/ - How to connect: Built-in n8n Linear node and trigger; Zapier and Make apps; Linear GraphQL API; Webhooks; TypeScript SDK; Linear MCP server - Authentication: Personal API key for internal scripts, or OAuth 2.0 for apps, including app-user installs for agents - Notes: An API key acts as the user who created it. OAuth apps can install as their own app user, so issues and comments show the integration rather than a person as the author. - Last updated: 2026-10-14 ## What you can automate with Linear - **Support-to-bug triage.** Turn escalated support conversations into Linear issues with reproduction steps, affected customers and a suggested team and priority. Guide: https://www.lumeor.studio/guides/ai-ticket-triage-and-routing/ - **Customer feedback to issues.** Cluster feature requests from support, sales calls and surveys and attach them to existing issues or projects with quotes and counts. Guide: https://www.lumeor.studio/guides/ai-customer-feedback-analysis/ - **Alert summaries.** Collapse related production alerts into one issue with an AI summary of what broke, where and since when. Guide: https://www.lumeor.studio/guides/ai-log-and-alert-summaries/ - **Project updates.** Draft weekly project updates from issue progress, merged pull requests and blockers for the project lead to edit and post. Guide: https://www.lumeor.studio/guides/automate-project-status-updates/ - **Meeting action items.** Create issues from decisions in planning and incident review meetings, linked to the right project and cycle. Guide: https://www.lumeor.studio/guides/ai-meeting-summaries-and-action-items/ - **Release notes drafting.** Collect completed issues in a cycle or release and draft customer-facing release notes grouped by theme. ## What makes Linear easy to automate Linear's model is opinionated: workspaces, teams, issues, cycles, projects, labels and workflow states, with fewer custom configurations than Jira. The API is GraphQL, the same one Linear's own apps use, so most things you can do in the UI are available programmatically. n8n, Zapier and Make all have Linear integrations, the official TypeScript SDK covers the schema, and webhooks are signed with an HMAC signature. Linear also treats agents as first-class. An OAuth app can install as its own app user, receive assignments and post activity in issues, and Linear provides an MCP server for assistants that need read and write access. That makes Linear one of the simpler places to plug in an agent that actually does work. ## The AI layer Linear is where product and engineering teams decide what to build next. AI is most useful at the boundaries: - **Inbound triage** from support and monitoring, detailed in [AI ticket triage and routing](/guides/ai-ticket-triage-and-routing/). - **Feedback synthesis**, linking customer requests to existing issues instead of creating duplicates, as in [analyzing customer feedback with AI](/guides/ai-customer-feedback-analysis/). - **Delegated agents** that pick up well-scoped issues and report progress in the thread. Before building an agent, check whether a packaged agent integration already covers the job. When it does not, the choice of framework matters; see [OpenAI Agents SDK vs Claude Agent SDK](/compare/openai-agents-sdk-vs-claude-agent-sdk/). ## Limits to plan around - **Hourly request and complexity budgets** that differ between API keys and OAuth apps. Deeply nested queries burn complexity quickly. - **Pagination.** Lists use cursor-based pagination; fetch in pages rather than requesting large nested collections. - **Webhook verification.** Validate the Linear-Signature header and the timestamp to reject replays. - **Workflow states per team.** Each team has its own states. Map them rather than assuming names. ## How Lumeor typically connects it For triage and feedback flows we use the n8n Linear node and Trigger with an OAuth app installed as its own user, an LLM step that searches existing issues before creating new ones, and validation against team, label and state IDs. For agent work we build a small service on the GraphQL API that receives assignments and posts updates back to the issue. ## Frequently asked questions ### What are Linear's API rate limits? Personal API keys allow 2,500 requests and 3,000,000 complexity points per hour per user. OAuth apps allow 5,000 requests and 2,000,000 complexity points per hour per user or app user. Requesting fewer fields and paginating sensibly keeps complexity down. ### Can an AI agent work inside Linear? Yes. Linear lets OAuth apps install as app users that can be mentioned, assigned or delegated issues and can post comments and updates. That makes it practical to run a coding or triage agent that takes work from Linear and reports back in the issue thread. ### Does n8n support Linear? Yes. n8n has a built-in Linear node for issues and comments and a Linear Trigger node for webhook events. For queries the node does not cover, the HTTP Request node can send GraphQL directly with the same credential. ### Linear or Jira for an automated workflow? Linear has a cleaner, more opinionated model and a modern API, so integrations are quicker to build. Jira is far more configurable and has broader enterprise and service desk features. Pick based on how your team works; both automate well. --- Source: https://www.lumeor.studio/tools/linear/ Published by Lumeor Studio, Newtown, PA. hello@lumeor.studio. Get a written scope: https://www.lumeor.studio/start/