timesfm
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.
The prediction MCP — score your prompt before you generate, so you never waste a credit.
教 AI 做这类工作的 Skill,可与本项一起使用
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.
Open-source intelligence for the global theater. Track everything from the corporate/private jets of the wealthy, and spy satellites, to seismic events in one unified interface. Hook an AI agent up to have it parse through data and find previously unseen correlations. The knowledge is available to…
scrape data from Google Maps. Extracts data such as the name, address, phone number, website URL, rating, reviews number, latitude and longitude, reviews,email and more for each place
列表按使用分(满分 100)排序,由以下部分组成
| 热度 | 12/80 | GitHub 星标或每周下载量:较大的计 60,较小的计 20 |
|---|---|---|
| 新鲜度 | 3/12 | 30 天内有发布或代码更新得满分 |
| 仍在响应 | 0/8 | 上次检查时云端端点有响应 |
| 使用分 | 15/100 |
质量: 基础检查通过 3/4 项(仅展示,不用于排序) · 上升势头: 暂无数据 (只用于“上升中”标签)
选择你使用的 AI 应用,然后复制设置。
用一句话让 AI 帮你安装
请帮我安装 Dali by Lulu。按照 https://mcp.popnx.com/i/lulu-the-narwhal-dali.md 操作,先把命令给我看,等我同意后再运行。粘贴到能在你电脑上工作的 AI(如 Claude Code、Cursor)。它会读取我们的步骤,先给你看命令,等你同意。无需密钥,无需注册。 · 查看 AI 会读取的全文
通过 3/4 项检查 · 在你的电脑上运行
第一步:你用哪个 AI?
选择你用的应用,步骤会显示在下方。
可连接 Claude,但需要先准备
需要先在电脑上安装 Python (uv)(只需一次) · 仅限电脑上的 Claude 应用(Windows 或 Mac),网页和手机不可用
把这段话复制到 Claude 的对话里。AI 会先检查电脑,告诉你要改哪个文件,并且每次都等你同意。
请在这台电脑上为 Claude 安装 MCP server Dali by Lulu,按顺序:1) 用 uv --version 检查是否有 uv。没有的话,请让我从 https://docs.astral.sh/uv/getting-started/installation/ 安装,不要自己安装。2) 只使用这份配置(已通过 PopNx MCP 检查):{ "mcpServers": { "dali": { "command": "uvx", "args": [ "dali-mcp" ] } } } 3) 告诉我要修改的文件完整路径或要运行的命令,等我同意。4) 不需要密钥。5) 按约定执行,然后确认 Dali by Lulu 的工具真的出现了。这段话不包含你的密钥。如果需要密钥,AI 会请你自己粘贴到文件或应用的输入框里。
已经装好了吗?先检查 · 打开终端,输入这条命令后按 Enter
uv --versionMac:按 ⌘ + Space,输入 Terminal 后按 Enter
Windows:打开“开始”菜单,输入 Windows PowerShell,然后选“打开”
还没安装: uv 没有点选式安装程序。从 uv 官方页面复制适合你系统(macOS 或 Windows)的一行命令,粘贴到终端 · 不需要单独安装 Python,uv 会在需要时自动下载
装好后关掉终端再打开,重新运行同一条命令,应显示版本号: uv --version
安装一次,所有需要 uv 的工具都能用
SettingsDeveloperEdit Config
{
"mcpServers": {
"dali": {
"command": "uvx",
"args": [
"dali-mcp"
]
}
}
}用 Dali by Lulu 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Claude 说正在使用 Dali by Lulu,或回答中出现 Dali by Lulu 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Anthropic 帮助页面核实 · 你的界面可能因版本略有不同
ChatGPT 暂时无法连接 Dali by Lulu
Dali by Lulu 需要安装在你的电脑上,而 ChatGPT 只能连接在网上运行的服务器
可连接 Claude Code,但需要先准备
需要先在电脑上安装 Python (uv)(只需一次)
把这段话复制到 Claude Code 的对话里。AI 会先检查电脑,告诉你要改哪个文件,并且每次都等你同意。
请在这台电脑上为 Claude Code 安装 MCP server Dali by Lulu,按顺序:1) 用 uv --version 检查是否有 uv。没有的话,请让我从 https://docs.astral.sh/uv/getting-started/installation/ 安装,不要自己安装。2) 只使用这份配置(已通过 PopNx MCP 检查):claude mcp add --scope user dali -- uvx dali-mcp 3) 告诉我要修改的文件完整路径或要运行的命令,等我同意。4) 不需要密钥。5) 按约定执行,然后确认 Dali by Lulu 的工具真的出现了。这段话不包含你的密钥。如果需要密钥,AI 会请你自己粘贴到文件或应用的输入框里。
已经装好了吗?先检查 · 打开终端,输入这条命令后按 Enter
uv --versionMac:按 ⌘ + Space,输入 Terminal 后按 Enter
Windows:打开“开始”菜单,输入 Windows PowerShell,然后选“打开”
还没安装: uv 没有点选式安装程序。从 uv 官方页面复制适合你系统(macOS 或 Windows)的一行命令,粘贴到终端 · 不需要单独安装 Python,uv 会在需要时自动下载
装好后关掉终端再打开,重新运行同一条命令,应显示版本号: uv --version
安装一次,所有需要 uv 的工具都能用
Mac:按 Command+空格,输入 Terminal · Windows:打开 PowerShell
claude mcp add --scope user dali -- uvx dali-mcp应能看到 Dali by Lulu 在列表中并显示已连接
用 Dali by Lulu 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Claude Code 说正在使用 Dali by Lulu,或回答中出现 Dali by Lulu 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Anthropic 帮助页面核实 · 你的界面可能因版本略有不同
可连接 Cursor,但需要先准备
需要先在电脑上安装 Python (uv)(只需一次)
把这段话复制到 Cursor 的对话里。AI 会先检查电脑,告诉你要改哪个文件,并且每次都等你同意。
请在这台电脑上为 Cursor 安装 MCP server Dali by Lulu,按顺序:1) 用 uv --version 检查是否有 uv。没有的话,请让我从 https://docs.astral.sh/uv/getting-started/installation/ 安装,不要自己安装。2) 只使用这份配置(已通过 PopNx MCP 检查):{ "mcpServers": { "dali": { "command": "uvx", "args": [ "dali-mcp" ] } } } 3) 告诉我要修改的文件完整路径或要运行的命令,等我同意。4) 不需要密钥。5) 按约定执行,然后确认 Dali by Lulu 的工具真的出现了。这段话不包含你的密钥。如果需要密钥,AI 会请你自己粘贴到文件或应用的输入框里。
已经装好了吗?先检查 · 打开终端,输入这条命令后按 Enter
uv --versionMac:按 ⌘ + Space,输入 Terminal 后按 Enter
Windows:打开“开始”菜单,输入 Windows PowerShell,然后选“打开”
还没安装: uv 没有点选式安装程序。从 uv 官方页面复制适合你系统(macOS 或 Windows)的一行命令,粘贴到终端 · 不需要单独安装 Python,uv 会在需要时自动下载
装好后关掉终端再打开,重新运行同一条命令,应显示版本号: uv --version
安装一次,所有需要 uv 的工具都能用
用 Dali by Lulu 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Cursor 说正在使用 Dali by Lulu,或回答中出现 Dali by Lulu 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Cursor 帮助页面核实 · 你的界面可能因版本略有不同
可连接 VS Code,但需要先准备
需要先在电脑上安装 Python (uv)(只需一次)
把这段话复制到 VS Code 的对话里。AI 会先检查电脑,告诉你要改哪个文件,并且每次都等你同意。
请在这台电脑上为 VS Code 安装 MCP server Dali by Lulu,按顺序:1) 用 uv --version 检查是否有 uv。没有的话,请让我从 https://docs.astral.sh/uv/getting-started/installation/ 安装,不要自己安装。2) 只使用这份配置(已通过 PopNx MCP 检查):{ "servers": { "dali": { "type": "stdio", "command": "uvx", "args": [ "dali-mcp" ] } } } 3) 告诉我要修改的文件完整路径或要运行的命令,等我同意。4) 不需要密钥。5) 按约定执行,然后确认 Dali by Lulu 的工具真的出现了。这段话不包含你的密钥。如果需要密钥,AI 会请你自己粘贴到文件或应用的输入框里。
已经装好了吗?先检查 · 打开终端,输入这条命令后按 Enter
uv --versionMac:按 ⌘ + Space,输入 Terminal 后按 Enter
Windows:打开“开始”菜单,输入 Windows PowerShell,然后选“打开”
还没安装: uv 没有点选式安装程序。从 uv 官方页面复制适合你系统(macOS 或 Windows)的一行命令,粘贴到终端 · 不需要单独安装 Python,uv 会在需要时自动下载
装好后关掉终端再打开,重新运行同一条命令,应显示版本号: uv --version
安装一次,所有需要 uv 的工具都能用
用 Dali by Lulu 告诉我你能帮我做什么,然后做一件给我看✓ 如果 VS Code 说正在使用 Dali by Lulu,或回答中出现 Dali by Lulu 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 VS Code 帮助页面核实 · 你的界面可能因版本略有不同
Claude Code
claude mcp add --scope user dali -- uvx dali-mcpClaude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"dali": {
"command": "uvx",
"args": [
"dali-mcp"
]
}
}
}Cursor (~/.cursor/mcp.json)
{
"mcpServers": {
"dali": {
"command": "uvx",
"args": [
"dali-mcp"
]
}
}
}VS Code (mcp.json)
{
"servers": {
"dali": {
"type": "stdio",
"command": "uvx",
"args": [
"dali-mcp"
]
}
}
}这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
依据这个服务器的连接方式和各应用自己的文档,检查于 2026年11月10日。
通过互联网连接,适用于
ClaudeGooseLM StudioMsty StudioCherry StudioCursorVS Code + GitHub CopilotVisual StudioWindsurf (Devin Desktop)ZedKiroClineContinueJetBrains AI AssistantClaude CodeCodex CLIGemini CLIOpenCodenodecodeWarpLibreChatOpen WebUIRaycast
在你的电脑上运行,适用于
ClaudeGooseLM StudioMsty StudioCherry StudioCursorVS Code + GitHub CopilotVisual StudioWindsurf (Devin Desktop)ZedKiroClineContinueJetBrains AI AssistantClaude CodeCodex CLIGemini CLIOpenCodenodecodeWarpLibreChatRaycast
我们上次检查时服务器没有响应,所以无法读取它的工具。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月19日
注册名称已通过 GitHub 账户 Lulu-The-Narwhal 验证,该账户拥有此仓库
代码最后更新于4个月前
使用 MIT
我们检查时,6 秒内没有响应
我们检查时,6 秒内没有响应
尚未检查;我们会在检查端点时读取工具列表
云端地址最近 5 次检查中有 0 次响应 · 最近一次没有响应 · 检查于 2026年11月19日
OSV.dev 中没有 dali-mcp@0.6.1 的记录
只检查这个版本的包本身,不包括所有依赖 · 检查于 2026年11月15日
如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/servers/lulu-the-narwhal-dali)提问、分享技巧或报告无法使用。按“有帮助”数量排序。
还没有评价
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 4 项适用的基础检查中通过了 3 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
两种方式都可以:安装在你的电脑上,或者连接云端版本。
GitHub 仓库最后更新于4个月前,最新版本(0.6.1)于4个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,因此也能用于接受远程 MCP 的应用,例如网页版 ChatGPT(Business 及以上为完整 MCP;Pro 为只读)。
同一分类中有:OpenSEO, Exa, Microsoft Learn MCP, Docs, COTAL actions (cotal.ai)
同一分类、其他作者、命令已通过检查。官方的排在前面,再按星标排序。
如何看卡片:↓ 每周下载量i★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Keyword research, site audits, rank tracking and AI visibility for your AI agent.
Connect AI agents to Exa for web search, content fetching, and multi-step research.
Official Microsoft Learn MCP Server – real-time, trusted docs & code samples for AI and LLMs.
cotal.ai actions: product overview, site search, build log, feedback, Cloud waitlist, updates, calls
| 本项 Dali by Lulu | OpenSEO | Exa | Microsoft Learn MCP | Docs | COTAL actions (cotal.ai) | |
|---|---|---|---|---|---|---|
| 运行位置 | 云端或安装在电脑上都可以 | 在云端运行,无需安装 | 在云端运行,无需安装 | 在云端运行,无需安装 | 在云端运行,无需安装 | 在云端运行,无需安装 |
| 需要密钥 | 否 | 否 | 否 | 否 | 否 | 否 |
| 可在 Claude 中按 + 添加 | 否 | 是 | 是 | 是 | 否 | 否 |
| 基础检查 | 3/4 | 5/5 | 2/4 | 4/4 | 4/4 | 4/4 |
| 许可证 | MIT | MIT | MIT | CC-BY-4.0 | NOASSERTION | Apache-2.0 |
| 价格 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 |
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
dali.getlulu.dev · Install · Live stats · Lulu
Score your creative against what's actually winning in the ad market — before you spend the credit.
Most AI generation failures are predictable. A weak prompt, an off-formula creative — you can't tell until after you've burned the token. Dali scores it *first*, and it doesn't grade against opinions or generic "prompt tips." It grades against a real, living corpus of proven-winning ads — creatives still running in the market months after launch, scraped, embedded, and ranked. Two jobs:
Every wasted generation has a real cost — a Seedance retry is ~$6. The live dashboard tracks what the community has saved by catching bad creatives before they burned a credit.
You: "make a video ad for our glass serum bottle"
dali::score_prompt(prompt, "veo3") → 8/100 Grade: F → no camera move · no motion · no lighting · 8 words → Verdict: Generic stock footage guaranteed. Enhance first. → enhancement_brief included (score < 70):
① lead with camera — Veo 3's #1 lever: "Slow dolly", "Orbital push" ② describe physics: "a drop falls", "liquid ripples", "glass refracts" ③ lighting type + quality: "warm backlight", "rim-lit edges" ↳ [Camera]. [Subject + motion]. [Lighting]. [Mood]. [No text.]
✦ YOUR LLM rewrites using the brief:
"Slow orbital push around a glass serum bottle on white marble. A single amber drop falls in extreme slow motion, catching warm backlight. Macro: liquid gold ripples outward from impact. Rim-lit edges, soft studio diffusion. Premium, clinical. No text."
dali::score_prompt(enhanced, "veo3") → 91/100 Grade: A ✓ Safe to generate.
This is what makes Dali more than a prompt linter. The scores are grounded in real ads that are actually winning, not hand-written rules.
How the corpus is built — longevity is the outcome signal. We scrape the public Meta Ad Library. An ad still running *months* after it launched is one the advertiser keeps paying for — a proven winner. That "still-running-after-N-days" longevity is a market-validated label you can't fake, and it's the spine of the whole dataset.
What's in it, today:
| Ads ingested | 10,204 (14,100 raw archive) |
| Proven winners (long-running) | 3,808 |
| Distinct advertisers | 4,121 |
| Verticals | 8 — beauty, wellness, supplements, fitness, food, apparel, tech, pets |
| Winner creatives embedded | 800 (1408-dim, balanced ~100/vertical) |
| Longest-running winner seen | 2,431 days (6.6 years live) |
The pipeline (offline → serving). The tools never scrape or embed on the fly — they read pre-built stores:
scrape Meta Ad Library → proven winners (longevity label)
→ Gemini vision → creative attributes (lighting, format, before/after, offer…) → prevalence SQL → winning-pattern lift per vertical (winners vs baseline) → Vertex embeddings → BigQuery VECTOR_SEARCH (nearest proven winners, cosine) → graph edges (Memgraph) → (:Pattern)-[:WINS_IN {lift, n}]->(:Category)
So when score_creative runs, it embeds your image and finds the actual winning ads it most resembles by full visual signature — then tells you which winning attributes you're missing. When enhance_prompt runs with a category, the rewrite brief is backed by real market lift ("before/after shows up in 78% of winning wellness ads, 4× baseline"), not craft opinion.
Honest scope. The winner label is *longevity* (a strong market-validated proxy), not per-ad conversion rate — measured CVR validation is in progress. The corpus grows on a schedule, so coverage per vertical keeps deepening. What you get today: your creative scored against what's *demonstrably surviving* in the live market.
dali::score_creative(image_url, "beauty")
→ score 62/100 — partial resemblance to proven winners → looks_like: Frøya Organics (ran 411d), tashportcosmetics (884d), Face Reality (346d) → what_to_change: winners use "before/after" 4× more · offer-visible 1.8× more → defects: none → Verdict: Partial — strong resemblance, but add the high-lift attributes before spending.
Hosted MCP — connect once, scores every prompt and creative:
# Claude Code
claude mcp add --transport http dali https://dali.getlulu.dev/mcp
// Cursor / Windsurf — .cursor/mcp.json or windsurf settings
{ "mcpServers": { "dali": { "url": "https://dali.getlulu.dev/mcp" } } }
// stdio-only clients — npx wrapper around the hosted server, no Python needed
{ "mcpServers": { "dali": { "command": "npx", "args": ["-y", "dali-mcp"] } } }
→ [Full install guide with all clients](https://dali.getlulu.dev/#install)
Self-hosted — local, no auth required:
pip install dali-mcp
claude mcp add dali -- python -m dali.server
The self-hosted package exposes the prompt-scoring tools locally. The creative-scoring tools (
score_creative,analyze_winning_formula) and the winning-ad corpus run on the hosted server — connect via the hosted MCP to use them.
Score the creative — against real winners
| Tool | What it does |
|---|---|
score_creative(image_url, category) | Score an actual ad image. Embedding similarity to proven winners is the headline score; also returns the winners it resembles, which winning attributes it's missing, and generation defects — in one call |
score_creative_from_view(category, …) | Score an image you're looking at (pasted/attached in the chat) — no URL. The model reads the creative's attributes and Dali scores them against the winning corpus (verdict + what to change). Use for images shared in-conversation; score_creative (URL) adds the embedding headline |
analyze_winning_formula(csv, category, email) | Paste your own ads export (creative URL + CPA/CTR/ROAS) → your winning formula vs your losers, plus how you compare to the industry median |
Score the prompt — before you generate
| Tool | What it does |
|---|---|
score_prompt(prompt, model, category?) | Grade 0–100 with a per-dimension breakdown and verdict. When the score is weak, the rewrite brief is returned in the same call. Reads intent with the conversation LLM (understands negation, any language) |
enhance_prompt(prompt, model, category?) | Returns a structured rewrite brief — YOUR LLM writes the enhanced prompt. With a category, the brief is backed by real winning-ad lift |
track_enhancement(original, enhanced, generator) | Record a before/after pair in the graph brain — trains community patterns |
score_variations(prompts, generator) | Rank a list of prompt variants in one call — highest to lowest |
suggest_generator(concept, budget_usd_max) | Pick the best model for your concept + budget |
The graph brain & meta
| Tool | What it does |
|---|---|
creative_patterns(model) | Community top patterns for this model from the graph |
community_benchmark(prompt, model) | Compare your prompt against community top scorers |
prompt_neighbors(prompt, model) | Find A/B-grade prompts that share your patterns (score the prompt first, so its patterns are in the graph) |
analyze_intent(prompt) | Parse dimensions: camera, motion, lighting, style, mood, gaps |
my_story() | Your scoring history, model stats, grade distribution |
list_generators() | All supported models with medium and core strength |
dali_version() | Server version + changelog |
| Model | Platforms | Best for | Prompt style |
|---|---|---|---|
veo3 | Higgsfield, Google AI Studio (veo-3.1-generate-preview), Runway | Cinematic brand films, narrative ads, photorealistic motion | Camera move → Subject → Action → Location → Lighting → Mood |
seedance | Higgsfield, fal.ai (bytedance/seedance-2.0) | UGC, social-native content, TikTok/Reels performance ads | Natural language, motion-first, authentic feel |
kling | Higgsfield (kling3), Kling.ai (kling-v3-text-to-video) | Character animation, product showcases, facial performance | Scene → Characters → Action → Camera → Style; multi-shot labels |
runway | Runway (gen4_turbo) | VFX, character performance, cinematic motion | Motion-first — describe what moves, not what exists |
wan | fal.ai (fal-ai/wan/v2.7/text-to-video) | 4K, 20-second clips, native audio, open-source workflows | Scene → Motion → Sound → Duration → Mood |
minimax | fal.ai (fal-ai/minimax/hailuo-02/pro/text-to-video) | Cinematic storytelling, character animation | Natural language + [camera movement] bracket syntax |
higgsfield | Higgsfield (native model) | Physics-driven motion — cloth, hair, fluid, particles | Describe materials in motion, not motion abstractly |
Sora 2 (OpenAI): API shutdown September 24, 2026. Do not build new dependencies on it — use Runway or Kling instead.
| Model | Platforms | Best for | Prompt style |
|---|---|---|---|
flux | BFL API (flux-pro-v1.1), fal.ai, Replicate | Photorealism, technical photography, product shots | 30–80 words; camera body + lens specs; front-load subject |
midjourney | Midjourney (v8.1) | Artistic depth, editorial, stylized illustration | Prose + params appended: --ar 16:9 --s 300 --v 8.1 --style raw |
| ideogram | Ideogram API (V_4), fal.ai | Typography,