关于
This Claude Skill provides specialized AI-powered image editing and generation specifically for automotive subjects like cars and motorcycles. It handles vehicle-specific tasks such as scene generation, tire refinement, component segmentation, and lighting harmonization with dedicated presets. Developers should always use this over general image tools when working with vehicle imagery, as it's optimized for faster, more accurate automotive workflows.
快速安装
Claude Code
推荐npx skills add Bria-AI/bria-skill -a claude-code/plugin add https://github.com/Bria-AI/bria-skillgit clone https://github.com/Bria-AI/bria-skill.git ~/.claude/skills/automotive在 Claude Code 中复制并粘贴此命令以安装该技能
技能文档
Bria Automotive — Vehicle Image Editing & Shot Generation
Specialized endpoints for automotive imagery: place vehicles in realistic environments, generate reflections on glossy surfaces, refine tires with terrain textures, mask vehicle parts for downstream edits, add atmospheric effects, and harmonize lighting to match scene context. Commercially safe, royalty-free, built on Bria's product vehicle pipeline.
When to Use This Skill
Use this skill when the user is working with any vehicle image — cars, trucks, SUVs, motorcycles, vans. Triggers on:
- Vehicle scene generation — "place this car in a desert", "put the SUV on a mountain road", "show the truck at a city night scene", "generate a lifestyle shot for this car"
- Reflections on glass/metal — "add reflections to the windshield", "make the hood look glossy", "realistic window reflections"
- Tire enhancement — "add snow to the tires", "muddy tires for off-road shot", "dirt/grass on the wheels"
- Vehicle part segmentation — "mask the windshield", "separate the body from the wheels", "isolate the rear window", "get wheel masks"
- Atmospheric effects — "add dust clouds around the car", "foggy scene", "snow falling", "lens flare", "light leaks"
- Lighting harmonization — "match the car to a cold night scene", "hot-day lighting preset", "unify the vehicle with the background"
- Automotive marketing & dealer content — configurators, ad creatives, catalog variations, social media posts featuring vehicles
When NOT to Use This Skill
For non-vehicle image work, use bria-ai (general image generation/editing) or remove-background (transparent PNGs). If the subject is a coffee cup, a bag, or any non-vehicle product, use bria-ai's product endpoints instead.
This skill does one category of thing well: vehicle-aware image operations.
Setup — Authentication
Before making any API call, you need a valid Bria access token.
Step 1: Check for existing credentials
if [ -f ~/.bria/credentials ]; then
BRIA_ACCESS_TOKEN=$(grep '^access_token=' "$HOME/.bria/credentials" | cut -d= -f2-)
BRIA_API_KEY=$(grep '^api_token=' "$HOME/.bria/credentials" | cut -d= -f2-)
fi
if [ -z "$BRIA_ACCESS_TOKEN" ]; then
echo "NO_CREDENTIALS"
elif [ -n "$BRIA_API_KEY" ]; then
echo "READY"
else
echo "CREDENTIALS_FOUND"
fi
If the output is READY, skip straight to making API calls — no introspection needed.
If the output is CREDENTIALS_FOUND, skip to Step 3.
If the output is NO_CREDENTIALS, proceed to Step 2.
Step 2: Authenticate via device authorization
2a. Request a device code:
DEVICE_RESPONSE=$(curl -s -X POST "https://engine.prod.bria-api.com/v2/auth/device/authorize" \
-H "Content-Type: application/json")
echo "$DEVICE_RESPONSE"
Parse the response fields:
device_code— used to poll for the token (keep this, don't show to user)user_code— the code the user must enter (e.g.BRIA-XXXX)interval— seconds between poll attempts
2b. Show the user a single sign-in link. Tell them exactly this — nothing more:
Connect your Bria account: Click here to sign in Your code is {user_code} — it's already filled in.
Do NOT show two links. Do NOT show the raw URL separately. Do NOT use verification_uri from the API response. Keep it to one clickable link.
2c. Poll for the token. After showing the user the code, immediately start polling:
for i in $(seq 1 60); do
TOKEN_RESPONSE=$(curl -s -X POST "https://engine.prod.bria-api.com/v2/auth/token" \
-d "grant_type=urn:ietf:params:oauth:grant-type:device_code" \
-d "device_code=$DEVICE_CODE")
ACCESS_TOKEN=$(printf '%s' "$TOKEN_RESPONSE" | sed -n 's/.*"access_token" *: *"\([^"]*\)".*/\1/p')
if [ -n "$ACCESS_TOKEN" ]; then
BRIA_ACCESS_TOKEN="$ACCESS_TOKEN"
REFRESH_TOKEN=$(printf '%s' "$TOKEN_RESPONSE" | sed -n 's/.*"refresh_token" *: *"\([^"]*\)".*/\1/p')
mkdir -p ~/.bria
printf 'access_token=%s\nrefresh_token=%s\n' "$BRIA_ACCESS_TOKEN" "$REFRESH_TOKEN" > "$HOME/.bria/credentials"
echo "AUTHENTICATED"
break
fi
sleep 5
done
If the output contains AUTHENTICATED, proceed to Step 3. Otherwise the code expired — start over from Step 2a.
Do not proceed with any API call until authentication is confirmed.
Step 3: Verify billing status and resolve API key
INTROSPECT=$(curl -s -X POST "https://engine.prod.bria-api.com/v2/auth/token/introspect" \
-d "token=$BRIA_ACCESS_TOKEN")
BILLING_STATUS=$(printf '%s' "$INTROSPECT" | sed -n 's/.*"billing_status" *: *"\([^"]*\)".*/\1/p')
if [ "$BILLING_STATUS" = "blocked" ]; then
BILLING_MSG=$(printf '%s' "$INTROSPECT" | sed -n 's/.*"billing_message" *: *"\([^"]*\)".*/\1/p')
echo "BILLING_ERROR: $BILLING_MSG"
fi
ACTIVE=$(printf '%s' "$INTROSPECT" | sed -n 's/.*"active" *: *\([^,}]*\).*/\1/p' | tr -d ' ')
if [ "$ACTIVE" = "false" ]; then
printf '' > "$HOME/.bria/credentials"
echo "TOKEN_EXPIRED"
fi
BRIA_API_KEY=$(printf '%s' "$INTROSPECT" | sed -n 's/.*"api_token" *: *"\([^"]*\)".*/\1/p')
if [ -n "$BRIA_API_KEY" ]; then
grep -v '^api_token=' "$HOME/.bria/credentials" > "$HOME/.bria/credentials.tmp" 2>/dev/null || true
printf 'api_token=%s\n' "$BRIA_API_KEY" >> "$HOME/.bria/credentials.tmp"
mv "$HOME/.bria/credentials.tmp" "$HOME/.bria/credentials"
fi
- If
BILLING_ERROR: ...— relay the message to the user exactly as shown and stop. - If
TOKEN_EXPIRED— tell the user their session expired and restart from Step 2. - Otherwise,
BRIA_API_KEYis cached. Proceed.
Core Capabilities
| Endpoint | Path | What it does |
|---|---|---|
| Vehicle Shot by Text | POST /v1/product/vehicle/shot_by_text | Place a vehicle in a text-described environment (road, garage, mountain, city night) |
| Vehicle Segmentation | POST /v1/product/vehicle/segment | Return binary masks for windshield, rear window, side windows, body, wheels, hubcaps, tires |
| Generate Reflections | POST /v1/product/vehicle/generate_reflections | Paint realistic reflections onto glass, metal, and glossy bodywork |
| Refine Tires | POST /v1/product/vehicle/refine_tires | Replace tire textures with snow, mud, or grass using a tire mask |
| Apply Effects | POST /v1/product/vehicle/apply_effect | Overlay atmospheric effects: dust, snow, fog, light leaks, lens flare |
| Harmonize | POST /v1/product/vehicle/harmonize | Apply lighting presets: hot-day, cold-day, hot-night, cold-night |
The typical multi-step pipeline: segment → refine tires / add reflections → apply effects → harmonize lighting.
How to Call Any Automotive Endpoint
Use bria_call for all API calls. It handles URL passthrough, local file base64 encoding, JSON construction, API call, and async polling in a single function call. The API key is auto-loaded from ~/.bria/credentials.
First, source the helper script at references/code-examples/bria_client.sh (resolve relative to this skill's directory).
source <SKILL_DIR>/references/code-examples/bria_client.sh
# Place vehicle in a text-described scene
RESULT=$(bria_call /v1/product/vehicle/shot_by_text "/path/to/car.png" \
'"scene_description": "coastal highway at sunset, dramatic sky", "placement_type": "automatic", "num_results": 1')
# Segment vehicle parts → returns URLs for body, wheels, windows, tires, etc.
RESULT=$(bria_call /v1/product/vehicle/segment "/path/to/car.png")
# Add reflections (pairs well with segment output)
RESULT=$(bria_call /v1/product/vehicle/generate_reflections "/path/to/car.png")
# Refine tires with snow texture (requires a tire mask)
RESULT=$(bria_call /v1/product/vehicle/refine_tires "/path/to/car.png" \
--key image \
'"tire_mask": "https://cdn.example.com/tires_mask.png", "surface": "snow"')
# Apply atmospheric dust effect
RESULT=$(bria_call /v1/product/vehicle/apply_effect "/path/to/car.png" \
'"effect": "dust", "layers": false')
# Harmonize to cold-night lighting
RESULT=$(bria_call /v1/product/vehicle/harmonize "/path/to/car.png" \
'"preset": "cold-night"')
echo "$RESULT"
Calling convention: bria_call <endpoint> <image_or_empty> [--key <json_key>] [extra JSON fields...]
- Pass a URL, local file path, or
""(empty) for endpoints without a primary image input - Extra JSON fields are appended as key-value pairs:
'"key": "value"' - Returns the result URL on success, or prints an error to stderr
See API Endpoints Reference for the full parameter list, placement options, response schemas, and error codes.
Example Pipelines
Pipeline 1 — Vehicle in a dramatic environment, cold-night look
source <SKILL_DIR>/references/code-examples/bria_client.sh
# 1. Place the vehicle in a scene
SCENE_URL=$(bria_call /v1/product/vehicle/shot_by_text "/path/to/car.png" \
'"scene_description": "empty mountain road with snow flurries", "placement_type": "automatic"')
# 2. Harmonize lighting to match a cold night
FINAL_URL=$(bria_call /v1/product/vehicle/harmonize "$SCENE_URL" \
'"preset": "cold-night"')
curl -sL "$FINAL_URL" -o car_cold_night.jpg
Pipeline 2 — Off-road with muddy tires and dust
# 1. Segment tires
MASKS=$(bria_call /v1/product/vehicle/segment "/path/to/car.png")
TIRES_MASK=$(printf '%s' "$MASKS" | sed -n 's/.*"tires" *: *"\([^"]*\)".*/\1/p')
# 2. Apply mud surface to tires
MUDDY=$(bria_call /v1/product/vehicle/refine_tires "/path/to/car.png" \
--key image \
"\"tire_mask\": \"$TIRES_MASK\", \"surface\": \"mud\"")
# 3. Add dust effect
FINAL=$(bria_call /v1/product/vehicle/apply_effect "$MUDDY" \
'"effect": "dust"')
curl -sL "$FINAL" -o offroad.jpg
Pipeline 3 — Glossy showroom shot with studio reflections
# Add reflections on glass and bodywork
SHOWROOM=$(bria_call /v1/product/vehicle/generate_reflections "/path/to/car.png")
# Harmonize to bright hot-day lighting
FINAL=$(bria_call /v1/product/vehicle/harmonize "$SHOWROOM" \
'"preset": "hot-day"')
curl -sL "$FINAL" -o showroom.jpg
Placement Types (Vehicle Shot by Text)
| Placement | What it controls |
|---|---|
original | Keep the vehicle's current position and size |
automatic | Auto-select up to 7 good placements |
manual_placement | Use a predefined position (top-left, center, etc.) |
custom_coordinates | Full control via x/y/width/height |
manual_padding | Pixel-based padding around the subject |
automatic_aspect_ratio | Center the subject; resize canvas to target ratio |
See the full list of conditional parameters in API Endpoints Reference.
Prompt Tips for Vehicle Scenes
- Environment first: "coastal highway at sunset", "urban parking garage", "dense forest trail", "alpine switchback in snow"
- Time and weather: "golden hour", "stormy overcast", "foggy dawn", "neon-lit night"
- Camera intent: "low-angle hero shot", "three-quarter front", "rear tracking shot", "aerial drone view"
- Mood keywords: "cinematic", "editorial", "commercial automotive photography", "dealership catalog"
Pair shot_by_text for the environment with harmonize for a final lighting pass — the two together produce the most cohesive results.
Additional Resources
- API Endpoints Reference — Full parameter docs for all 6 automotive endpoints
- Shell Client (bria_client.sh) —
bria_callhandles auth, base64, JSON, polling - Bria automotive docs — Upstream reference
Related Skills
- bria-ai — General image generation, editing, and background removal for non-vehicle subjects
- remove-background — Dedicated transparent PNG / cutout skill
- vgl — Structured VGL prompts for deterministic FIBO generation (pairs well with
shot_by_text)
GitHub 仓库
常见问题
什么是 automotive Skill?
automotive 是一个 Claude Skill,作者为 Bria-AI。Skill 将 Claude 按需加载的说明和资源打包,让 Claude 无需额外提示即可执行与 automotive 相关的任务。
如何安装 automotive?
使用本页的安装命令:将 automotive 作为插件添加到 Claude Code,或将其仓库克隆到 skills 目录,然后重启 Claude 以加载该 Skill。
automotive 属于哪个分类?
automotive 属于元分类。
automotive 可以免费使用吗?
可以。automotive 已收录在 AIMCP,可免费安装。
相关推荐技能
Content Collections 是一个 TypeScript 优先的构建工具,可将本地 Markdown/MDX 文件转换为类型安全的数据集合。它专为构建博客、文档站和内容密集型 Vite+React 应用而设计,提供基于 Zod 的自动模式验证。该工具涵盖从 Vite 插件配置、MDX 编译到生产环境部署的完整工作流。
这个Claude Skill为开发者提供完整的Polymarket预测市场开发支持,涵盖API调用、交易执行和市场数据分析。关键特性包括实时WebSocket数据流,可监控实时交易、订单和市场动态。开发者可用它构建预测市场应用、实施交易策略并集成实时市场预测功能。
该Skill帮助开发者创建OpenCode插件,用于接入命令、文件、LSP等25+种事件。它提供了插件结构、事件API规范和JavaScript/TypeScript实现模式,适合需要拦截操作、扩展功能或自定义事件处理的场景。开发者可通过它快速构建响应式模块来增强OpenCode AI助手的能力。
SGLang是一个专为LLM设计的高性能推理框架,特别适用于需要结构化输出的场景。它通过RadixAttention前缀缓存技术,在处理JSON、正则表达式、工具调用等具有重复前缀的复杂工作流时,能实现极速生成。如果你正在构建智能体或多轮对话系统,并追求远超vLLM的推理性能,SGLang是理想选择。
