Fal.ai MCP

為什麼要用

核心特性

Image: Flux, SDXL, Stable Diffusion 3, Ideogram, Recraft
Video: Kling, Hailuo, Veo, AnimateDiff
Audio/Music: MusicGen, Stable Audio, MMAudio
TTS: ElevenLabs, F5-TTS, Kokoro
Per-call pricing with the Fal balance shown in tool output

即時演示

實際使用效果

fal-mcp.replay ▶ 就緒

0/0

安裝

選擇你的客戶端

~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "fal-mcp": {
      "command": "uvx",
      "args": [
        "fal-mcp-server"
      ]
    }
  }
}

開啟 Claude Desktop → Settings → Developer → Edit Config。儲存後重啟應用。

~/.cursor/mcp.json · .cursor/mcp.json

{
  "mcpServers": {
    "fal-mcp": {
      "command": "uvx",
      "args": [
        "fal-mcp-server"
      ]
    }
  }
}

Cursor 使用與 Claude Desktop 相同的 mcpServers 格式。專案級設定優先於全域。

VS Code → Cline → MCP Servers → Edit

{
  "mcpServers": {
    "fal-mcp": {
      "command": "uvx",
      "args": [
        "fal-mcp-server"
      ]
    }
  }
}

點擊 Cline 側欄中的 MCP Servers 圖示，然後選 "Edit Configuration"。

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "fal-mcp": {
      "command": "uvx",
      "args": [
        "fal-mcp-server"
      ]
    }
  }
}

格式與 Claude Desktop 相同。重啟 Windsurf 生效。

~/.continue/config.json

{
  "mcpServers": [
    {
      "name": "fal-mcp",
      "command": "uvx",
      "args": [
        "fal-mcp-server"
      ]
    }
  ]
}

Continue 使用伺服器物件陣列，而非映射。

~/.config/zed/settings.json

{
  "context_servers": {
    "fal-mcp": {
      "command": {
        "path": "uvx",
        "args": [
          "fal-mcp-server"
        ]
      }
    }
  }
}

加入 context_servers。Zed 儲存後熱重載。

claude mcp add fal-mcp -- uvx fal-mcp-server

一行命令搞定。用 claude mcp list 驗證，claude mcp remove 移除。

使用場景

實戰用法： Fal.ai MCP

Turn a script into a narrated short with one prompt

👤 Marketers, indie creators ⏱ ~15 min intermediate

何時使用： You have a 30-second script and want a complete narrated short.

前置條件

Server/skill installed and authenticated — See repo README

步驟

Generate the scenes

Split this 30s script into 4 scenes and generate a video clip for each with Kling.✓ 已複製

→ Four video URLs
Narrate it

TTS each scene's narration with ElevenLabs Rachel voice.✓ 已複製

→ Audio URLs aligned with scenes

結果： Four video clips + matching narration audio, ready to stitch.

注意事項

Costs add up fast — Costs add up fast — set a hard budget in the prompt: 'Use the cheapest model that meets quality bar' and check the balance reported in tool output.

搭配使用： filesystem

組合

與其他 MCP 搭配，撬動十倍槓桿

fal-mcp + filesystem

Pair with filesystem for complementary capabilities

Use this server together with filesystem to complete a multi-step task.✓ 已複製

工具

此 MCP 暴露的能力

工具	輸入參數	何時呼叫	成本
generate_image	prompt, model, aspect_ratio	Need a still image	$0.003–0.05 per image
generate_video	prompt, model, duration_s	Need a short video clip	$0.20–2.00 per clip
tts	text, voice, model	Produce narration audio	$0.0001 per char

成本與限制

運行它的成本

API 配額: See provider docs for rate limits
每次呼叫 Token 數: Varies by tool
費用: See repo README for pricing details
提示: Cache tool results and avoid repeated identical calls.

安全

權限、密鑰、影響範圍

憑證儲存： Use environment variables; never commit secrets

資料出站： Tool calls go to the provider's API as documented

故障排查

常見錯誤與修復

FAL_KEY missing

Get a key at fal.ai/dashboard/keys and set FAL_KEY env var.

驗證： Run generate_image once with a tiny prompt

insufficient balance

Top up at fal.ai/dashboard/billing. The error includes the required amount.

驗證： Check fal.ai/dashboard

替代方案

Fal.ai MCP 對比其他方案

替代方案	何時用它替代	權衡
Replicate MCP	You want different model selection	Different per-model pricing; similar feature set