deepseek-ai/awesome-deepseek-agent
1
0
Fork 0
mirror of https://github.com/deepseek-ai/awesome-deepseek-agent.git synced 2026-11-05 01:07:37 +00:00
Code Issues 354 Projects Packages Wiki Activity

[PR #445] docs: add bilingual Trae DeepSeek integration guide #441

New issue
Open
opened 2026-10-26 09:49:26 +00:00 by bsakaguchi · 0 comments
bsakaguchi commented 2026-10-26 09:49:26 +00:00
Owner
Copy link

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/awesome-deepseek-agent/pull/445
Author: @Shutong-Hou
Created: 10/9/2026
Status: 🔄 Open

Base: main ← Head: main

📝 Commits (3)

📊 Changes

4 files changed (+100 additions, -0 deletions)

View changed files

📝 README.md (+1 -0)
📝 README.zh-CN.md (+1 -0)
➕ docs/trae.md (+49 -0)
➕ docs/trae.zh-CN.md (+49 -0)

📄 Description

Summary

Add English and Simplified Chinese integration guides for using DeepSeek in Trae, a desktop AI IDE by ByteDance. Both guides follow the existing docs/ structure - installation, configuration, first run, and troubleshooting - and are submitted together with alphabetically placed README entries.

Files:

  • docs/trae.md + docs/trae.zh-CN.md
  • One row added to README.md and README.zh-CN.md (between Reasonix and WorkBuddy/CodeBuddy)

Scope: documentation only. No images, no code changes.

Which approach the guide covers

Trae ships DeepSeek-V4-Flash as a built-in model, billed against the Trae plan. This guide instead documents the preset provider route, where the reader selects DeepSeek as the provider in Settings > Models and enters their own DeepSeek API key. Calling a custom model does not consume the Trae plan allowance; usage is billed to the reader's DeepSeek account.

Critical checks

  1. Model naming - only deepseek-v4-pro and deepseek-v4-flash appear. No deprecated V3 names.
  2. 1M context - DeepSeek V4's 1M context is stated, and the guide tells readers to click the one-click 1M input chip. For output it only recommends leaving the field empty or using the 128k chip, since saving a manually entered larger value has not been verified. The 384K maximum output is mentioned in prose as a model capability, not as a recommended setting.
  3. Max thinking - DeepSeek V4 supports reasoning_effort up to max on the API side. Trae does not expose a per-level effort selector: its model panel exposes Thinking Mode as Follow model default / Enabled / Disabled. The guide documents the switch Trae actually provides and advises keeping it on follow-default or enabled, rather than disabling thinking to work around an API error. No workaround that degrades the model is proposed.
  4. Pricing - no price table is included, to avoid drift. The guide links the official DeepSeek pricing page.
  5. Config fields - every field name, label and option in the guide was checked against Trae's official models documentation and verified in the Trae model panel.

Note on overlap with the Qoder entry

Trae (ByteDance) and the existing Qoder entry (Alibaba Cloud) are separate products with different DeepSeek integration surfaces. Qoder is a three-form agentic coding product (IDE / CLI / JetBrains plugin) whose Custom tab configures a model provider; Trae is a desktop AI IDE whose DeepSeek integration runs through the preset-provider panel in Settings > Models. The endpoints, fields and steps differ.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/deepseek-ai/awesome-deepseek-agent/pull/445 **Author:** [@Shutong-Hou](https://github.com/Shutong-Hou) **Created:** 10/9/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `main` --- ### 📝 Commits (3) - [`ece0fa1`](https://github.com/deepseek-ai/awesome-deepseek-agent/commit/ece0fa10adeac32cd066f3ed0818b5afe245bd9b) Add files via upload - [`adaa098`](https://github.com/deepseek-ai/awesome-deepseek-agent/commit/adaa0983d6ba1104383161626fa573ff5d4ee98b) Update README.md - [`5685433`](https://github.com/deepseek-ai/awesome-deepseek-agent/commit/56854338a3c6c85f137019ba78823daa71ab06d7) Update README.zh-CN.md ### 📊 Changes **4 files changed** (+100 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `README.md` (+1 -0) 📝 `README.zh-CN.md` (+1 -0) ➕ `docs/trae.md` (+49 -0) ➕ `docs/trae.zh-CN.md` (+49 -0) </details> ### 📄 Description ## Summary Add English and Simplified Chinese integration guides for using DeepSeek in **Trae**, a desktop AI IDE by ByteDance. Both guides follow the existing `docs/` structure - installation, configuration, first run, and troubleshooting - and are submitted together with alphabetically placed README entries. Files: - `docs/trae.md` + `docs/trae.zh-CN.md` - One row added to `README.md` and `README.zh-CN.md` (between Reasonix and WorkBuddy/CodeBuddy) Scope: documentation only. No images, no code changes. ## Which approach the guide covers Trae ships DeepSeek-V4-Flash as a built-in model, billed against the Trae plan. This guide instead documents the **preset provider** route, where the reader selects DeepSeek as the provider in `Settings > Models` and enters their own DeepSeek API key. Calling a custom model does not consume the Trae plan allowance; usage is billed to the reader's DeepSeek account. ## Critical checks 1. **Model naming** - only `deepseek-v4-pro` and `deepseek-v4-flash` appear. No deprecated V3 names. 2. **1M context** - DeepSeek V4's 1M context is stated, and the guide tells readers to click the one-click `1M` input chip. For output it only recommends leaving the field empty or using the `128k` chip, since saving a manually entered larger value has not been verified. The 384K maximum output is mentioned in prose as a model capability, not as a recommended setting. 3. **Max thinking** - DeepSeek V4 supports `reasoning_effort` up to `max` on the API side. Trae does not expose a per-level effort selector: its model panel exposes Thinking Mode as Follow model default / Enabled / Disabled. The guide documents the switch Trae actually provides and advises keeping it on follow-default or enabled, rather than disabling thinking to work around an API error. No workaround that degrades the model is proposed. 4. **Pricing** - no price table is included, to avoid drift. The guide links the official DeepSeek pricing page. 5. **Config fields** - every field name, label and option in the guide was checked against Trae's official models documentation and verified in the Trae model panel. ## Note on overlap with the Qoder entry Trae (ByteDance) and the existing Qoder entry (Alibaba Cloud) are separate products with different DeepSeek integration surfaces. Qoder is a three-form agentic coding product (IDE / CLI / JetBrains plugin) whose Custom tab configures a model provider; Trae is a desktop AI IDE whose DeepSeek integration runs through the preset-provider panel in `Settings > Models`. The endpoints, fields and steps differ. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
No results found.
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/awesome-deepseek-agent#441
No description provided.