What is the difference between IDE mode and SOLO mode?

IDE mode keeps the editor, terminal, debugging, extensions and source control, so every change stays under manual control. In SOLO mode the AI takes the lead: you state the requirement in natural language or by voice, and it plans the task and works through code generation, testing, preview and a change summary.

Can TRAE use my own model or API key?

Yes. Alongside the built-in models, a custom model can be added with your own API key, using either the OpenAI Chat Completions or the Anthropic Messages API format, with a preset or fully custom request URL.

How is usage measured?

Each paid plan carries a dollar usage quota that is drawn down per request, sized to match the monthly price: $8 on Lite, $20 on Pro, $60 on Pro+ and $200 on Ultra. The free plan has a limited monthly allowance instead.

Does TRAE support MCP servers?

Yes. MCP servers can be added from a marketplace or configured manually as JSON, over stdio or HTTP, and an existing mcp.json from another IDE can be reused. Individual agents can be given their own set of MCP servers.

TRAE

An AI development environment with a hands-on IDE mode and a hands-off SOLO mode.

Starting price: Free tier; Pro $20 / month
Free trial: Free plan with a limited monthly allowance
Output: AI-assisted editing and autonomous task execution in an IDE

Visit TRAE

Overview

TRAE is a development tool built around two ways of working with AI on the same project. IDE mode keeps the familiar editor, terminal, debugger and source control, so changes stay under manual control; SOLO mode hands the task over, taking a requirement in natural language or by voice and planning, generating, testing and previewing the result on its own. The inline assistant, CUE, covers completion, multi-line edits and next-edit prediction. Several models are built in, and further ones can be added with an API key.

Our take

The two modes are the reason to look at this rather than a single-mode assistant: the same project can be driven line by line one hour and handed over whole the next, without moving to a different tool when the task changes shape. Bringing your own API key keeps that flexible, since the plan quota and the model bill can be separated when a team already buys tokens elsewhere. The pricing needs reading closely, though. Usage is a dollar quota drawn down per request rather than a request count, so a month spent on an expensive model costs more than the same number of calls on a cheap one, and the cheapest paid plan is sold only in a dozen countries, which puts most buyers at the twenty-dollar tier or above. Anyone who needs a documented API to drive the tool from a build pipeline should look elsewhere: this is an editor to sit in, not a service to call.

Key Facts

Category	AI Code
Starting price	Free tier; Pro $20 / month Annual billing brings Pro to $200 per year, about $16.70 per month. Pro+ $60 and Ultra $200 monthly, or $600 and $2,000 annually. A cheaper Lite plan at $8 per month is sold only in a listed set of countries. Each paid plan carries a dollar usage quota of the same size as its monthly price.
Free trial	Free plan with a limited monthly allowance
Output	AI-assisted editing and autonomous task execution in an IDE
Hosting	Desktop IDE, plus TraeWork on web, desktop and mobile

Screenshots

Placeholders. They are replaced as soon as the vendor sends material, and both open the same short form.

Features

Where it runs: Its own desktop IDE with two modes, IDE and SOLO, plus TraeWork as a separate web, desktop and mobile client; remote work over Remote SSH and WSL
Inline completions: Yes, CUE: code completion, chained completion, multi-line edits, next-edit prediction and navigation, plus dependency imports and reference renaming in Python, TypeScript and Go
Agent mode: Yes, agents that plan multi-step work and call tools, with custom agents carrying their own prompts, MCP servers and toolsets; SOLO mode runs planning, generation, testing and preview end to end from a spoken or written requirement
Model choice: Yes, built-in models include GPT-6-Astra, GPT-5.6-Sol, GPT-5.6-Terra, GPT-5.6-Luna, GPT-5.5, GPT-5.4, GPT-5.2, Gemini-3.1-Pro-Preview, Gemini-3-Flash-Preview, Seed-2.1-Turbo, GLM-5.2, MiniMax-M3, MiniMax-M2.7 and Kimi-K2.5; custom models can be added with an API key over the OpenAI Chat Completions or Anthropic Messages format. The free plan is limited to automatic model selection, paid plans allow manual choice
Data control: Privacy mode stops chats, code snippets and AI output being used for data analysis, product optimisation or model training, and codebase files stay on the local device; agent commands can be run in a sandbox with file access controls and high-risk command blocking
Ecosystem fit: Extension store, built-in Git workflow with AI-written commit messages, AI code review over uncommitted changes, single commits or branch diffs, and MCP servers added from a marketplace or by pasting JSON
API: No public API for driving the product; enterprise deployments are arranged through BytePlus

Integrations

Figma
Playwright

Deployment and Support

Deployment: Cloud (SaaS)
Browser-based
Support: Email helpdesk
Onboarding: Documentation and knowledge base
Company size: Freelancers
Small and mid-sized
Enterprise

Suitability

Industry	Freelancers	Small and mid-sized	Enterprise
Agencies	Not stated	Not stated	Not stated
Architecture & Engineering	Not stated	Not stated	Not stated
Consulting	Not stated	Not stated	Not stated
Media & Creative	Not stated	Not stated	Not stated
Technology	Stated as a fit	Stated as a fit	Stated as a fit

The table combines the industries the vendor addresses with the company sizes the tool is aimed at. An empty cell means the combination is not stated, not that it is ruled out.

Pros and Cons

Strengths

Two modes in one tool: keep manual control in IDE mode or hand the whole task over in SOLO mode
Custom models can be added with your own API key over OpenAI- or Anthropic-compatible endpoints
Privacy mode keeps chats and code out of training, and agent commands can run sandboxed
Free plan with a monthly allowance, no time limit

Limitations

Usage is metered as a dollar quota drawn down per request, so cost depends on which model a task uses
The cheapest paid plan, Lite, is sold only in a listed set of countries
The free plan cannot choose a model manually and has limited autocompletion
No public API for driving the tool from outside

Pricing

Free

$0

Limited monthly allowance, automatic model selection only, standard queue, limited autocompletion and 2 concurrent cloud tasks.
Lite

$8

Per month, or $80 a year. An $8 monthly usage quota, manual model selection, fast queue, unlimited autocompletion and 2 concurrent cloud tasks. Sold only in Thailand, Indonesia, Vietnam, the Philippines, India, Pakistan, Turkey, Malaysia, Brazil, Colombia, Peru and Nigeria.
Pro

$20

Per month, or $200 a year (about $16.70 a month). A $20 monthly usage quota and 10 concurrent cloud tasks.
Pro+

$60

Per month, or $600 a year (about $50 a month). A $60 monthly usage quota and 15 concurrent cloud tasks.
Ultra

$200

Per month, or $2,000 a year (about $166.70 a month). A $200 monthly usage quota, 20 concurrent cloud tasks and early access to new models.

Best For

Technology

Alternatives and Comparisons

FAQ

What is the difference between IDE mode and SOLO mode?: IDE mode keeps the editor, terminal, debugging, extensions and source control, so every change stays under manual control. In SOLO mode the AI takes the lead: you state the requirement in natural language or by voice, and it plans the task and works through code generation, testing, preview and a change summary.
Can TRAE use my own model or API key?: Yes. Alongside the built-in models, a custom model can be added with your own API key, using either the OpenAI Chat Completions or the Anthropic Messages API format, with a preset or fully custom request URL.
How is usage measured?: Each paid plan carries a dollar usage quota that is drawn down per request, sized to match the monthly price: $8 on Lite, $20 on Pro, $60 on Pro+ and $200 on Ultra. The free plan has a limited monthly allowance instead.
Does TRAE support MCP servers?: Yes. MCP servers can be added from a marketplace or configured manually as JSON, over stdio or HTTP, and an existing mcp.json from another IDE can be reused. Individual agents can be given their own set of MCP servers.

What users say

Reviews

0 reviews

Ease of use: Not rated yet
User interface: Not rated yet
Onboarding: Not rated yet
Support and service: Not rated yet
Value for money: Not rated yet
Features: Not rated yet

No reviews yet. Nothing has been published here yet, and no score is shown. How a review is checked before it appears is described below.

How we check reviews

Reviews are not published yet. The form on a tool's page works and reaches us, but nothing from it appears on this site, so no score here is based on a review. The rules below are the ones that will apply once reviews go live.

Every review is read by a person before it appears. We require a working email address and use it to come back to the author, so a review has a reachable sender rather than being anonymous. We publish positive and negative reviews alike, we pay nothing and offer nothing for a review, and we accept none that was bought or incentivised.

What we cannot do is verify a purchase. There is no shop here and no account, so whether the author really used the tool rests on what they tell us, plus the plausibility of what they write. We say so rather than implying a check that does not exist.

A vendor cannot have a review removed, cannot buy one, and cannot have one changed. Paid placement and a paid listing have no effect on which reviews appear or on any score.

Share your experience

Write a Review

Ready to take a closer look at TRAE

The link goes straight to the vendor. Some links on this site are affiliate links.