经 AI Skill Hub 精选评估，AI个人助手 获评「强烈推荐」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色，AI 评分 8.0 分，适合有一定技术背景的用户使用。
AI个人助手 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排，将复杂的多步骤任务拆解为清晰的自动化流程，实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成，适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
AI个人助手 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排，将复杂的多步骤任务拆解为清晰的自动化流程，实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成，适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一：npm 全局安装 npm install -g mf0-1984 # 方式二：npx 直接运行（无需安装） npx mf0-1984 --help # 方式三：项目依赖安装 npm install mf0-1984 # 方式四：从源码运行 git clone https://github.com/PavelMuntyan/MF0-1984 cd MF0-1984 npm install npm start
# 命令行使用
mf0-1984 --help
# 基本用法
mf0-1984 [options] <input>
# Node.js 代码中使用
const mf0_1984 = require('mf0-1984');
const result = await mf0_1984.run(options);
console.log(result);
# mf0-1984 配置说明 # 查看配置选项 mf0-1984 --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量（覆盖配置文件） export MF0_1984_CONFIG="/path/to/config.yml"
MF0.ai is a multi-provider AI chat platform — a single interface for ChatGPT, Perplexity, Gemini, Claude, Grok, and Seed, with a persistent Memory tree (3D knowledge graph) with semantic search, themes (folders with per-theme rules) and dialogs, analytics (with MF0 platform fee), assistant favorites, bug reporting (with optional screenshot attachment), project profile backup/restore (.mf bundles), Audiochat — real-time voice chat optimised for hands-free use, a Gallery panel (Profile → Gallery) showing all AI-generated images and videos with lightbox navigation, deep chat links, and per-tile download buttons, Fusion (Слияние) — multi-agent parallel deep-research mode that queries all four providers simultaneously and synthesises one unified document, a Pricing panel (Profile → Pricing) with the full in-app tariff for all chat, image, and video services, message deletion (X button on each chat bubble permanently removes the message and any attached files), an adaptive composer (prompt field shrinks to half height in active chats to show more conversation), and a built-in Help assistant (Profile → Help) that answers product questions using this repo's HANDOFF.md as its knowledge base.
Audiochat streams voice via the OpenAI Realtime and Grok Realtime APIs. Sessions are stateless (no history, no Memory Tree) and are saved to a dedicated "Аудиочаты" folder, with each session named by model and date (e.g. Grok-2026-05-09-23-15). The UI uses large tap targets so the feature is comfortable to use while driving.
| **UI dev server** | Vite — default port **1984** (vite.config.js) |
| **Local API** | Express 5 + Node.js — default port **35184** (API_PORT) |
| **Database** | SQLite (default) or Postgres (DB_ADAPTER=postgres) |
| **Auth** | Optional OAuth (Google / GitHub) when DB_ADAPTER=postgres |
| **Version** | **2.8.5** (package.json) — see HANDOFF.md for release notes |
For architecture, data model, and operations, see HANDOFF.md (engineering handoff).
---
npm install
npm run devThe dev setup runs the API and Vite together. Open the URL Vite prints (typically http://127.0.0.1:1984).
API health: GET http://127.0.0.1:35184/api/health — expect JSON with "ok": true and "mfLabApi": true.
---
Plugins add AI generation capabilities beyond text chat. All generation is fully asynchronous — users can queue multiple generations simultaneously without waiting.
Pricing (+ 10% MF0 fee):
| Model | Price per image |
|---|---|
| Nano Banana Pro | $0.04 |
| Nano Banana 2 | $0.04 |
| Flux 2 | $0.04 |
| Seedream 4.5 | $0.04 |
| Qwen 2 Pro | $0.04 |
Takes a source image + text prompt → generates a short video clip. Pricing (+ 10% MF0 fee, per second of output):
| Model | Price/sec | 5 sec | 10 sec | 15 sec |
|---|---|---|---|---|
| Kling 3 | $0.168 | $0.84 | $1.68 | $2.52 |
| Seedance 2 | $0.3024 | $1.51 | $3.02 | $4.54 |
| Veo 3.1 | $0.20 | — | 4s=$0.80 | 6s=$1.20 / 8s=$1.60 |
| Wan 2.7 | $0.10 | $0.50 | $1.00 | $1.50 |
Async generation: Submit → server queues the generation job → polls every 3 sec → client polls /api/fal/job/:turnId every 5 sec. On failure, balance is refunded automatically. Multiple concurrent generations across different models are fully supported.
| Model | Inputs | Effect | Price |
|---|---|---|---|
| Kling 3 | image + video | Motion transfer: character from image, motion from video | $0.168/sec |
| Veed | video + audio | Lipsync — video loops to match audio length | $0.40/min |
| Sync Lipsync 2 | video + audio | Lipsync — output length = source video length | $5.00/min |
| Topaz | video | AI upscaling (720p=$0.01/s, 1080p=$0.02/s, >1080p=$0.08/s) | variable |
Cost charged post-completion from actual output duration. Min balance $5 to show plugin.
---
高质量的AI工作流项目，值得关注
该工具未明确声明开源协议，商业使用前请联系原作者确认授权范围，避免侵权风险。
AI Skill Hub 为第三方内容聚合平台，本页面信息基于公开数据整理，不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后，再部署至生产环境，并做好必要的安全评估。
AI Skill Hub 点评：AI个人助手 的核心功能完整，质量优秀。对于自动化工程师和运维人员来说，这是一个值得纳入个人工具库的选择。建议先在非生产环境试用，再逐步推广。
| 原始名称 | MF0-1984 |
| Topics | chatbot3d-visualizationjavascript |
| GitHub | https://github.com/PavelMuntyan/MF0-1984 |
| 语言 | JavaScript |
收录时间：2026-06-17 · 更新时间：2026-06-17 · License：未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型，复制安装指令后粘贴到对应客户端