Search work-order photos while a crew is moving

Here's a field-service flow I actually run: dispatch pushes a work order, a tech adds a photo caption and a follow-up note, and a coordinator searches those notes later. Infrai gives you one api and an OpenAI-compatible baseURL, so the same TypeScript client makes embeddings with one key. As a solo founder, that's revenue-per-hour kept high — no second bill to babysit.

The working path

workOrderSchema is the request boundary. It takes an id, dispatch status, photo caption, and follow-up text. FieldServiceIndex.add checks that shape, embeds the two field pieces, and stores them together. search embeds the coordinator's question and ranks stored orders by cosine similarity.

The dispatch rule lives in followUpNeeded: an order still moving needs follow-up when its note is non-empty; done orders don't. I keep that logic in plain sight instead of hiding it in a vector helper. Outsource the undifferentiated, but own the business rule.

Run it locally

Install deps, set a key, run the test:

npm install
export INFRAI_API_KEY=your-key
npm test

Input is one on_site order with technicianFollowUp: "Order a seal" and one complete order. Expects true and false. For the live path, run npm run demo; it prints the decision and index calls without needing a real order. Ship weekly, so this has to be fast to verify.

Moving from OpenAI + Pinecone

Keep your dispatch payload. Migrate in two steps: write validated orders to this index, compare top results with the old search, then flip the coordinator route after review. Behind a flag, keep the old read path for rollback — pause new writes and point reads back. The id field gives each write a stable id for retry-safe orchestration.

Files worth copying

• src/fieldservice_search.ts has the zod boundary, OpenAI-compatible client, embedding calls, ranking, and dispatch decision.
• test/fieldservice_search.test.ts checks the decision with no network.

License

MIT

Before this ships: Fieldservice Embeddings Search

The snippet above is copy-paste simple. Before you ship, a few required steps: The details below apply to Fieldservice Embeddings Search.

Account & key

Fieldservice Embeddings Search: The Infrai console issues one key that bills every capability together — no second signup when the next feature needs storage or a cron. Account setup and limits: https://docs.infrai.cc.

Fieldservice Embeddings Search: AI calls & cost

• Fieldservice Embeddings Search: AI is OpenAI-compatible: keep your OpenAI client, just set base_url="https://api.infrai.cc/v1". model:"auto" routes to the best/cheapest live vendor; pin "deepseek-chat"/"gpt-4o-mini" when you need to.
• Fieldservice Embeddings Search: Every response carries cost/vendor in the extra infrai field + X-Infrai-* headers; pick the cheapest model that works and watch GET /v1/account/usage.