Trae

ByteDance’s AI-first IDE, shipped as a VS Code fork with extensions and keybindings that transfer in one import. Flagship agent is SOLO mode, which plans, edits files across the project, and shows live preview without step-by-step approvals. The multi-model backbone covers Claude, OpenAI, DeepSeek, and Gemini routes on a single token-metered bill.

Pricing went from fully free through 2025 to token-based tiers in February 2026. Entry paid plan is $3/mo Lite, roughly an order of magnitude under Cursor and Windsurf. Free tier still exists with 5,000 autocompletes and 2 concurrent cloud tasks.

ByteDance’s wider AI stack (Doubao, Jimeng, Seedance, Kling) sits behind the same wallet. Trae is best read as a loss-leader play for developer mindshare, not a pure standalone product.

System Verdict

Pick Trae if you want the cheapest route to Claude or GPT inside a full IDE. Lite at $3/mo is the most aggressive paid coding tier in the market. SOLO mode handles greenfield agent runs competently, Figma-to-code works on screenshots, and the VS Code fork means existing extensions port over cleanly.

Skip it if you need the strongest terminal agent, or if ByteDance-adjacent telemetry is a dealbreaker. Claude Code still leads on CLI-native agentic coding. Cursor has the deeper IDE polish and enterprise governance. Security researchers have flagged telemetry collection even with settings disabled, which rules Trae out for regulated codebases.

Who pays which tier: Free for casual or learning use (5K autocompletes, limited models), Lite $3/mo for solo devs wanting Claude and GPT on a budget, Pro $10/mo for SOLO mode access, Pro+ $30/mo for heavier agent runs, Ultra $100/mo for power users who want 20x Pro usage and model early-access.

Key Facts

What it actually is

A VS Code fork with ByteDance’s AI layer bolted in at the editor level. Users import existing VS Code settings, themes, and extensions on first launch. The differentiator is the agent layer, not the editor chrome.

SOLO mode is the headline feature. It takes a natural-language spec, plans the steps, edits multiple files, runs terminal commands, and surfaces a live preview. It is closest in positioning to Cursor’s YOLO mode and Windsurf’s Cascade.

Builder mode is the lighter variant for smaller scoped tasks: breaks a request into sub-steps, modifies files, previews results.

Model routing runs across Claude, GPT, DeepSeek, and Gemini on the same token bill. Users pick per-request or let Trae auto-route. Doubao models are expected to join the IDE routing soon given the parent-company stack.

When to pick Trae

• Budget coding. $3/mo Lite is the cheapest verified path to Claude and GPT inside an IDE. Cursor Pro is $20/mo, Windsurf sits at $15/mo. For hobbyists or early-career devs, the spread matters.
• Autonomous agent runs on greenfield projects. SOLO mode plans and executes without the hand-holding you need in Copilot Chat or basic autocomplete.
• Design-to-code. The Figma screenshot import to React component is a first-party feature and works without an extra plugin.
• Existing VS Code users. Keybindings, themes, and extensions transfer. Muscle memory is preserved.
• Multi-model comparison inside one IDE. Pick Claude for reasoning, GPT for breadth, DeepSeek for cost, Gemini for long context, on a single subscription.

When to pick something else

• Strongest CLI agent: Claude Code. Terminal-native agentic coding across full codebases is still Anthropic’s league.
• Deep IDE polish and enterprise governance: Cursor. Longer track record, clearer data-handling docs, broader team features.
• Google Cloud-native workflow: Antigravity. First-party Google agentic IDE with Gemini integration.
• Open-source self-host: Cline or Continue. Run any model, keep all code local.
• Instant web-app prototyping: Bolt, Lovable, or v0. Browser-based, zero-setup, full-stack generation.
• Windsurf-style Cascade agent: Windsurf. More mature Cascade workflow, Codeium-backed.
• Privacy-first lightweight editor: Zed. Fast, minimal, AI-optional.

Pricing

Subscription pricing verified 2026-04-18 via trae.ai/pricing:

Pro includes a 7-day free trial. Lite had a $3 first-month promo historically; verify current offers on the pricing page.

Models available across paid tiers span Claude, OpenAI, DeepSeek, and Gemini routes. Free tier has limited model selection.

Usage is metered on a token pool per tier, not unlimited. This shifted in February 2026 from the prior fully-free Pro feature set.

Prices verified 2026-04-18 via Trae pricing page and ohaiknow Trae review 2026.

Against the alternatives

Failure modes

• Telemetry concerns. Security researchers have flagged continued data collection even with settings disabled. Regulated codebases should avoid.
• Token pool limits surprise users. The February 2026 shift from fully-free Pro to token-based tiers caught heavy users mid-workflow. Budget the month.
• ByteDance geopolitical exposure. US enterprise buyers in defence, finance, or government-adjacent work may be blocked by procurement policy regardless of the tool’s merits.
• Model versions lag frontier. Some currently verified routing targets lag current frontier model versions. Frontier users running Claude Opus 4.7 or OpenAI frontier models on direct subscriptions will notice the gap.
• SOLO mode drifts on complex codebases. Autonomous agents still over-edit unrelated files when the spec is vague. Keep scope tight on legacy projects.
• Support is thin compared to Cursor. Community docs exist, but first-party responsiveness has been uneven during the pricing transition.
• Cloud task limits are low on free and Lite. 2 concurrent tasks is a hard ceiling at the lower tiers, which bottlenecks parallel agent work.

Methodology

This page was produced by the aipedia.wiki editorial pipeline, an automated system that ingests vendor documentation, verifies pricing and model details against primary sources, and generates the editorial analysis you are reading. No individual human wrote this review. Scoring follows the four-dimension rubric at /about/scoring/ (Utility × Value × Moat × Longevity, unweighted average). Last verified 2026-04-18 against trae.ai/pricing and ohaiknow Trae review.

FAQ

Is Trae free? Yes. The free tier gives 5,000 autocompletes per month and 2 concurrent cloud tasks. Paid tiers start at $3/mo for Lite, which unlocks premium models.

Who owns Trae? ByteDance, the same parent company behind Doubao, Jimeng, Seedance, and Kling. TikTok is also ByteDance.

What models does Trae support? Claude, OpenAI, DeepSeek, and Gemini routes are the currently verified model families. Selection depends on plan tier.

What is SOLO mode? Trae’s autonomous agent. It plans a task, edits multiple files, runs terminal commands, and shows live preview without step-by-step approval. It is included on Pro and higher tiers.

How does Trae compare to Cursor? Trae is cheaper at every tier ($3 and $10 vs $20 for Cursor Pro) and routes to more models by default. Cursor has the longer track record, stronger enterprise governance, and deeper IDE polish. Full comparison at /tools/cursor/.

Can I run Trae on Linux? Yes. As a VS Code fork, it ships on Windows, macOS, and Linux.

Is my code sent to ByteDance servers? AI inference runs through ByteDance-routed infrastructure by default. Security researchers have flagged telemetry even with settings disabled. Do not use Trae on regulated or proprietary code without a clear data-handling review.

Related

• Category: AI Coding
• Alternatives: Cursor · Windsurf · Cline · Continue · Claude Code · Aider · Codeium · Zed · GitHub Copilot · Antigravity
• Browser-based app builders: Bolt · Lovable · v0 · Replit Agent

Embed this score on your site Free. Links back.

HTML Copy

<a href="https://aipedia.wiki/tools/trae/" target="_blank" rel="noopener"><img src="https://aipedia.wiki/badges/trae.svg" alt="Trae on aipedia.wiki" width="260" height="72" /></a>

Markdown Copy

[![Trae on aipedia.wiki](https://aipedia.wiki/badges/trae.svg)](https://aipedia.wiki/tools/trae/)

Badge value auto-updates if the editorial score changes. Attribution via the link is required.