Alpha

The Knowledge Platform for AI Agents

Keep your knowledge base fresh with automatic syncing. Search with semantic precision. Serve complete context to your agents. All in one platform built for production RAG systems.

Join the Waitlist→View Docs

agent.ts

import { streamText, tool } from 'ai';
import { openai } from '@ai-sdk/openai';
import { z } from 'zod';

const result = streamText({
  model: openai.responses('gpt-5'),
  tools: {
    'search-knowledge': tool({
      description: 'Search knowledge base, returns chunks',
      parameters: z.object({ query: z.string() }),
      execute: async ({ query }) => {
        const res = await fetch(
          `https://lupa.wtf/api/search?projectId=${projectId}&deploymentId=${deploymentId}&query=${query}`
        );
        return res.json();
      }
    }),
    'get-snapshot-contents': tool({
      description: 'Get full document content',
      parameters: z.object({ snapshotId: z.string() }),
      execute: async ({ snapshotId }) => {
        const res = await fetch(
          `https://lupa.wtf/api/snapshots/${snapshotId}`
        );
        return res.text();
      }
    })
  }
});

AI SDK compatible

Used by engineers at

Built for Builders Who've Been There

Stop fighting your knowledge base. Start building.

"Set it to refresh daily and forget about it. My docs are always up-to-date without any manual work."

— AI Engineer at YC Startup

"Search results come back in under 50ms. Our users actually notice the difference."

— Full-stack Developer, Enterprise SaaS

"I've built RAG systems before. This is the first one that didn't make me want to pull my hair out."

— Senior Engineer at AI Startup

Automatic sync

Schedule daily, weekly, or monthly refreshes for websites. Upload new file versions anytime. Changes are automatically detected, processed, and deployed. Your knowledge stays fresh without manual work.

Semantic search

Vector search returns ranked chunks with similarity scores in under 50ms. Find relevant content across your entire knowledge base. Perfect for discovery and quick answers.

Full document retrieval

Get complete markdown content of any document snapshot. Multi-step agents search to discover, then retrieve full context for comprehensive, accurate answers.

Production analytics

Monitor query patterns, document hits, zero-result searches, and performance metrics. Optimize your RAG pipeline with real-time insights powered by Tinybird.

Already syncing automatically. Coming soon: advanced filtering, metadata search, and folder-level operations

Complete workflow: Sync, Search, Serve

From automatic updates to intelligent retrieval. Your knowledge base works for you, not the other way around.

Knowledge stays fresh automatically

Set refresh schedules for websites or upload new file versions. Lupa detects changes, processes content, and updates your deployments without manual intervention.

Agents search with semantic precision

Search API returns top 5 semantic matches with scores in under 50ms. Agent identifies relevant documents by snapshotId for deeper analysis.

Full context drives accurate answers

When needed, agents fetch complete document content for comprehensive reasoning. With full context and OpenAI Responses, they deliver thorough, well-informed answers.

// Multi-step RAG with reasoning
import { streamText } from 'ai';
import { openai } from '@ai-sdk/openai';

const result = streamText({
  model: openai.responses('gpt-5'),
  providerOptions: {
    openai: {
      reasoningEffort: 'low',
      reasoningSummary: 'detailed'
    }
  },
  tools: {
    'search-knowledge': searchTool,
    'get-snapshot-contents': getSnapshotTool
  },
  stopWhen: stepCountIs(15)
});

// Agent workflow:
// 1. User asks question
// 2. Agent searches knowledge base
// 3. Agent sees snapshotId repeated in results
// 4. Agent retrieves full document content
// 5. Agent reasons over complete context
// 6. Agent provides comprehensive answer

Build smarter AI agents

Join engineering teams building production AI agents with Lupa

Join the Waitlist→View on GitHub