chore: updated to version v0.1.1

.github/workflows/commit-message.yml

@@ -0,0 +1,52 @@
+name: Commit Message
+
+on:
+  pull_request:
+    branches:
+      - main
+      - dev
+
+permissions:
+  contents: read
+
+jobs:
+  validate:
+    name: Validate commit messages
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Check commit message subjects
+        shell: bash
+        run: |
+          set -euo pipefail
+
+          pattern='^(build|chore|ci|docs|feat|fix|perf|refactor|revert|style|test): [A-Za-z0-9 _-]+$'
+
+          range="${{ github.event.pull_request.base.sha }}..${{ github.event.pull_request.head.sha }}"
+
+          invalid=0
+          while IFS= read -r line; do
+            sha="${line%% *}"
+            subject="${line#* }"
+
+            if [[ "$subject" == Merge\ * ]]; then
+              continue
+            fi
+
+            if [[ ! "$subject" =~ $pattern ]]; then
+              echo "::error title=Invalid commit message::${sha} ${subject}"
+              invalid=1
+            fi
+          done < <(git log --format='%H %s' "$range")
+
+          if [[ "$invalid" -ne 0 ]]; then
+            echo "Commit message subjects must use: type: subject"
+            echo "Allowed types: build, chore, ci, docs, feat, fix, perf, refactor, revert, style, test"
+            echo "Allowed subject characters: English letters, numbers, spaces, hyphens, and underscores"
+            exit 1
+          fi

.github/workflows/docker-image.yml

@@ -4,11 +4,13 @@ on:
   push:
     branches:
       - main
+      - dev
     tags:
       - "v*.*.*"
   pull_request:
     branches:
       - main
+      - dev
   workflow_dispatch:
 
 permissions:

CONTRIBUTING.md

@@ -36,6 +36,33 @@ Use the example configuration files for local development. Do not commit local s
 - Keep generated artifacts out of commits unless the project explicitly requires them.
 - Do not commit caches, build output, `.pyc` files, `.env` files, or local storage data.
 
+## Commit Messages
+
+Use the `type: subject` format for every commit message subject.
+Use only English letters, numbers, spaces, hyphens, and underscores in the subject.
+
+Allowed types:
+
+- `build`
+- `chore`
+- `ci`
+- `docs`
+- `feat`
+- `fix`
+- `perf`
+- `refactor`
+- `revert`
+- `style`
+- `test`
+
+Examples:
+
+- `feat: add model routing priority`
+- `fix: handle expired refresh tokens`
+- `refactor: simplify channel lookup`
+
+Commit message subjects that do not match this format will fail CI.
+
 ## Architecture Boundaries
 
 - `frontend/` owns the user interface, client-side state, message rendering, and admin/user workflows.

README.md

@@ -34,9 +34,9 @@ The architecture is designed for simple deployment, efficient static delivery, a
 | Area | Capabilities |
 | --- | --- |
 | Conversations | Multi-branch chat, streaming, retries, edits, feedback, sharing, cloned shared conversations, rich markdown rendering, file cards, model metadata, usage details, and execution traces. |
-| Media generation | Dedicated image generation and image edit flow with task-aware routing, OpenAI Images-compatible protocols, generated file storage, preview, download, and run history separated from text chat. |
+| Media generation | Dedicated image generation and image edit flow with task-aware routing, provider-native OpenAI, Google, and xAI image protocols, generated file storage, preview, download, and run history separated from text chat. |
 | Model control plane | Platform model catalog, upstream channels, real upstream models, route bindings, priority and weight routing, model capability JSON, display ordering, vendor mapping, automatic icons, and circuit breaker state. |
-| Provider protocols | OpenAI Responses and Chat Completions, OpenAI Images, Anthropic Messages, Google/Gemini Generate Content, xAI Responses, OpenRouter defaults, and custom OpenAI-compatible routes. |
+| Provider protocols | OpenAI Responses, Chat Completions, Images Generations, and Images Edits; Anthropic Messages; Google/Gemini Generate Content and Image Generation; xAI Responses, Images Generations, and Images Edits; OpenRouter defaults; and custom OpenAI-compatible routes. |
 | Request governance | Protocol-aware request assembly, user option allowlists and denylists, system-protected fields, previous-response continuation where supported, and context snapshots for review. |
 | Files and RAG | File upload, preview, download, deletion, quota control, MIME detection, text extraction, OCR, full-context injection, image context, chunking, embeddings, and semantic retrieval. |
 | Memory and context | Message-window truncation, token-budget truncation, context compression, conversation memory, long-term user memory, RAG evidence records, and prompt trace inspection. |
@@ -246,7 +246,9 @@ Common backend environment variables:
 
 Production mode rejects unsafe default secrets, weak encryption keys, wildcard CORS, and non-HTTPS public URLs.
 
-The initial superadmin bootstrap credentials are built in as `deeix-chat` / `deeix-chat-2026` / `System Admin`. They are used only when the database has no superadmin account. The first login forces changing the username and password; later changes are managed from the account flow, not from `config.yaml`.
+The initial superadmin username is `admin`. When the database has no superadmin account, the backend generates a random password and prints it once in the startup logs while creating the account. The first login forces changing the username and password; later changes are managed from the account flow, not from `config.yaml`.
+
+To retrieve the initial admin password, inspect the backend logs from the first startup and search for `bootstrap superadmin created`; the `username` and `password` fields are the initial login credentials. If a superadmin already exists in the database, the service does not regenerate or print this password again.
 
 ## Security Notes
 

README.zh-CN.md

@@ -34,9 +34,9 @@ DEEIX Chat 为团队提供统一的 AI 工作台，用一个清晰的使用入
 | 模块 | 能力 |
 | --- | --- |
 | 对话体验 | 多分支会话、流式响应、重试、编辑、反馈、公开分享、克隆分享会话、富文本 Markdown、文件卡片、模型元信息、用量明细和执行链路。 |
-| 媒体生成 | 独立的图片生成和图片编辑链路，按任务类型路由到 OpenAI Images 兼容协议，生成结果统一入库为文件，支持预览、下载和独立运行记录。 |
+| 媒体生成 | 独立的图片生成和图片编辑链路，按任务类型路由到 OpenAI、Google 和 xAI 的原生图片协议，生成结果统一入库为文件，支持预览、下载和独立运行记录。 |
 | 模型控制面 | 平台模型目录、上游渠道、真实上游模型、路由绑定、优先级/权重路由、能力 JSON、展示顺序、厂商映射、自动图标和熔断状态。 |
-| 协议适配 | OpenAI Responses 与 Chat Completions、OpenAI Images、Anthropic Messages、Google/Gemini Generate Content、xAI Responses、OpenRouter 默认协议和自定义 OpenAI 兼容路由。 |
+| 协议适配 | OpenAI Responses、Chat Completions、Images Generations 和 Images Edits，Anthropic Messages，Google/Gemini Generate Content 和 Image Generation，xAI Responses、Images Generations 和 Images Edits，OpenRouter 默认协议和自定义 OpenAI 兼容路由。 |
 | 请求治理 | 按协议组装上游请求，支持用户参数白名单/黑名单、系统保护字段、协议支持时的 previous response 续接，以及可回看的上下文快照。 |
 | 文件与 RAG | 文件上传、预览、下载、删除、配额控制、MIME 探测、文本提取、OCR、全文上下文注入、图片上下文、分片、向量嵌入和语义检索。 |
 | 记忆与上下文 | 消息数截断、Token 预算截断、上下文压缩、会话记忆、用户长期记忆、RAG 证据记录和提示词链路查看。 |
@@ -246,7 +246,9 @@ pnpm build
 
 生产模式会拒绝不安全的默认密钥、过短的加密密钥、通配 CORS 和非 HTTPS 公开地址。
 
-初始化超级管理员凭据内置为 `deeix-chat` / `deeix-chat-2026` / `System Admin`。它只在数据库中不存在超级管理员时用于首次创建账号。首次登录会强制修改用户名和密码；后续账号变更通过账户流程完成，不再通过 `config.yaml` 修改。
+初始化超级管理员用户名固定为 `admin`，密码会在数据库中不存在超级管理员时随机生成，并只在首次创建账号的后端启动日志中输出一次。首次登录会强制修改用户名和密码；后续账号变更通过账户流程完成，不再通过 `config.yaml` 修改。
+
+获取初始化管理员密码时，请查看首次启动后端服务的日志，搜索 `bootstrap superadmin created`；其中的 `username` 和 `password` 分别是初始登录用户名和密码。如果数据库中已经存在超级管理员，服务不会重新生成或再次输出该密码。
 
 ## 安全说明
 

VERSION

@@ -1 +1 @@
-0.1.0
+0.1.1

backend/cmd/server/main.go

@@ -9,7 +9,7 @@ import (
 )
 
 // @title DEEIX Chat API
-// @version 0.1.0
+// @version 0.1.1
 // @description DEEIX Chat 后端 API 文档
 // @BasePath /api/v1
 // @securityDefinitions.apikey BearerAuth