n8n 和 Make.com 工作流的微调工具调用
用微调本地模型替换 n8n 或 Make.com 工作流中的 OpenAI 节点。相同的工具路由、相同的结构化输出、零 API 费用。
您有一个运行中的 n8n 工作流。客户发送消息，OpenAI 节点决定采取什么操作，工作流相应分支。每天运行 500 次，OpenAI 账单 450 美元/月且在增长。
以下是完全消除该费用的模式：从现有工作流执行中提取训练数据，在特定工具 schema 上微调小型模型，使用 Ollama 本地部署，将 OpenAI 节点替换为指向 localhost 的 HTTP Request 节点。
相同工作流。相同路由逻辑。相同结构化输出。零 API 费用。
架构
当前：Webhook → OpenAI Node → Switch Node → Action Branches
替换：Webhook → HTTP Request (本地 Ollama) → Switch Node → Action Branches
唯一改变的节点是 AI 决策点。下游一切保持不变。
费用对比
| 日执行量 | OpenAI 费用（GPT-4o） | 微调本地 | 月节省 |
|---|---|---|---|
| 100 | $90/月 | $0/月 | $90/月 |
| 500 | $450/月 | $0/月 | $450/月 |
| 2,000 | $1,800/月 | $0/月 | $1,800/月 |
| 5,000 | $4,500/月 | $0/月 | $4,500/月 |
可靠性提升
微调 7B 模型在客户支持工作流中的测试结果：
- **工具选择准确率：**94-97%（vs GPT-4o-mini 的 89-93%）
- **参数提取准确率：**91-95%（vs 87-92%）
- **Schema 合规性：**99.5%+（vs 96-98%）
- **幻觉工具名称：**0%（vs 1-3%）
Ship AI that runs on your users' devices.
Ertas early bird pricing starts at $14.50/mo — locked in for life. Plans for builders and agencies.
延伸阅读
Ship AI that runs on your users' devices.
Early bird pricing starts at $14.50/mo — locked in for life. Plans for builders and agencies.
Keep reading
Stop Paying GPT-4 to Call Your APIs: Fine-Tune a Local Tool-Calling Model
You're paying frontier-model prices for what amounts to pattern matching and JSON generation. A fine-tuned 8B model handles tool calling at 90%+ accuracy for zero per-query cost. Here's the math and the migration path.
Building AI Agents That Work Offline: Fine-Tuned Models for Edge Automation
AI agents that depend on cloud APIs are fragile, expensive, and privacy-risky. Fine-tuned tool-calling models running on edge hardware create agents that work offline, respond instantly, and keep data local.
Building Reliable AI Agents with Fine-Tuned Local Models: Complete Guide
Most AI agents are just GPT-4 wrappers — expensive, unreliable at scale, and dependent on cloud APIs. Fine-tuned local models hit 98%+ accuracy on your specific tools at zero per-query cost. Here's the complete architecture.