What's new
in Axon

Axon agents can now run multiple tasks in parallel. Instead of waiting for one task to complete before starting the next, Axon spins up parallel agents to work on independent subtasks simultaneously.

What's new

• Agents now detect task dependencies automatically and parallelize where possible

• New task graph view shows what's running in parallel vs. sequentially

• Average task completion time reduced by up to 60% on large refactors

How to use it

No configuration required. Axon detects parallelizable tasks automatically. You can view the task graph in the agent panel by clicking View graph during any agent run.

Axon now supports Google's Gemini 2.5 Pro as a model option. Switch between Claude, GPT-4o, and Gemini 2.5 Pro directly from the model selector in the editor.

What's new

• Gemini 2.5 Pro available in all plans (Pro and Team)

• Model switching is now instant — no restart required

• Per-project model preferences saved automatically

Notes

Gemini 2.5 Pro performs particularly well on large context windows and multi-file refactors. We recommend trying it on tasks that span more than 10 files.

Codebase indexing is now up to 4× faster on large repositories. We rewrote the indexing pipeline to process files in parallel and cache incremental changes rather than re-indexing from scratch.

Improvements

• Initial index time reduced by ~75% on repos over 50k lines

• Incremental updates now happen in the background without interrupting suggestions

• Memory usage during indexing reduced by 40%

Impact

Users on large monorepos should notice significantly faster first-suggestion times after opening a project.

Axon can now review your pull requests directly from GitHub. Connect your GitHub account and Axon will automatically review new PRs, flag issues, and suggest improvements as inline comments.

What's new

• Automatic PR review on open and updated pull requests

• Inline comments with specific line references

• Configurable review depth: quick scan vs. deep review

• Available on Pro and Team plans

Setup

Go to Settings → Integrations → GitHub and connect your account. Select which repositories you want Axon to review automatically.

A bug caused Axon to produce malformed suggestions when working with complex TypeScript generics, particularly nested conditional types and mapped types.

Fixed

• Suggestions no longer break mid-token on nested generic types

• Conditional type inference now handled correctly

• Mapped type completions restored

Affected versions

This bug affected v2.1.0–v2.1.3. Update to v2.1.4 to resolve it.

Axon is now available as a native plugin for JetBrains IDEs, including IntelliJ IDEA, WebStorm, PyCharm, and GoLand.

What's new

• Full feature parity with the VS Code extension

• Native JetBrains UI integration — no separate window required

• Codebase indexing works across all JetBrains project types

• Available on all plans

Install

Search for Axon in the JetBrains Marketplace, or install directly from Settings → Plugins inside your JetBrains IDE.

Axon's v1 REST API will be deprecated on July 1, 2025. All integrations should migrate to the v2 API before that date.

What's changing

• v1 API endpoints will return 410 Gone after July 1, 2025

• v2 API is already available and stable

• All v1 functionality is available in v2 with improved performance

Migration

See the v2 API migration guide for step-by-step instructions. The main changes are authentication (now uses API keys instead of OAuth tokens) and response format (now returns JSON:API compliant responses).

Contact support if you need help migrating before the deadline.