AICore Labs
    Getting started

    Documentation

    Everything you need to take agents, models and automations to production with AICore. REST endpoints, typed SDKs, runnable examples and deployment best practices.

    182 ms
    Latency p50
    99.95%
    Uptime SLA
    TS · Py · Go
    SDKs

    Quickstart

    Install the SDK and make your first call in under 2 minutes.

    # npm
npm install @aicore/sdk

# python
pip install aicore
    bash
    import { AICore } from "@aicore/sdk";

const ai = new AICore({ apiKey: process.env.AICORE_KEY });

const res = await ai.chat.completions.create({
  model: "atlas-70b",
  messages: [{ role: "user", content: "Summarize this email" }],
});

console.log(res.choices[0].message.content);
    typescript
    Interactive

    Playground

    Try the models right from the docs. Pick a model, ask a question and stream the response.

    Try it live

    Prova una di queste domande:

    Authentication

    All requests require a Bearer token. Keys are managed from the workspace panel with per-environment scope (dev / staging / prod) and automatic rotation every 90 days.

    curl https://api.aicore.dev/v1/chat/completions \
  -H "Authorization: Bearer $AICORE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"atlas-70b","messages":[{"role":"user","content":"hello"}]}'
    bash

    REST API

    Base URL https://api.aicore.dev/v1. All endpoints support streaming via SSE, function calling and response_format JSON-schema.

    POST
    /chat/completions

    Chat completion (optional streaming).

    POST
    /embeddings

    Semantic embeddings 1024d/3072d.

    POST
    /agents/runs

    Run an agent with tool calling.

    GET
    /agents/runs/:id

    Status and trace of a run.

    POST
    /files

    Upload documents for RAG (PDF, MD, DOCX).

    POST
    /evals

    Run an eval suite.

    SDKs

    First-class official SDKs in TypeScript, Python and Go. Exponential retry, streaming, typed responses.

    TypeScript

    v2.4.1
    npm i @aicore/sdk

    Python

    v2.4.0
    pip install aicore

    Go

    v1.8.2
    go get aicore.dev/sdk

    Webhooks

    Receive real-time events for completed runs, failures and processed batches. HMAC-SHA256 signature.

    {
  "id": "evt_01J9Z7...",
  "type": "agent.run.completed",
  "data": {
    "run_id": "run_xyz",
    "agent": "sales-qualifier",
    "duration_ms": 4318,
    "tokens": { "in": 1240, "out": 410 }
  },
  "created_at": "2026-06-17T14:22:01Z"
}
    json

    Models

    atlas-70b

    Frontier reasoning, advanced function calling.

    ctx 128k
    helix-8b

    Fast and cheap for high-volume tasks.

    ctx 32k
    mosaic-vision

    Multimodal: documents, screenshots, OCR.

    ctx 16k
    ember-embed-3

    3072d embeddings for semantic retrieval.

    ctx —

    Agents

    Define agents with tool calling, persistent memory and guardrails. Run sandbox-isolated.

    const agent = ai.agents.define({
  name: "sales-qualifier",
  model: "atlas-70b",
  tools: [hubspot.getDeals, gmail.draftReply],
  guardrails: ["no-pii-leak", "tone-professional"],
});

const run = await agent.run({ input: "Qualify today's leads" });
    typescript

    Cookbooks

    RAG with verifiable citations

    Cookbook · ~15 min

    Sales agent on HubSpot

    Cookbook · ~15 min

    Document AI: invoice extraction

    Cookbook · ~15 min

    Evals with golden dataset

    Cookbook · ~15 min

    Fine-tuning on Helix 8B

    Cookbook · ~15 min

    Streaming with SSE in React

    Cookbook · ~15 min