AIOS
Agentic Operating System
by Swan AI Labs

Your Data. Plain English.
Any Tool.

Connect your databases and manage your entire data infrastructure in plain English — queries, table creation, stored procedures, ETL, everything — from any IDE, LLM, or AI agent.

10 database types — SQL, NoSQL, and cloud warehouses
IDEs Cursor • VS Code • Windsurf
LLMs Claude • ChatGPT • Gemini
CLIs Claude Code • Terminal • Scripts
Agents LangChain • CrewAI • AutoGen
10
Database Types
5 sec
To Connect
$0
To Start

Up and Running in Minutes

Four steps from zero to full data intelligence across your entire organization.

1

Sign Up & Get Your Key

Create a free account. You get an API key instantly — this is your agent's identity in the AIOS system.

Takes 30 seconds. No credit card.
2

Connect Your Databases

Paste a connection string. AIOS supports 10 database types — SQL Server, PostgreSQL, MySQL, MongoDB, Oracle, SQLite, Snowflake, BigQuery, Redshift, ClickHouse.

5 seconds per database. Read-only access is fine.
3

Connect Your IDE or LLM

Drop your AIOS key into Claude Code, Cursor, Windsurf, ChatGPT, or any tool that talks HTTP. One header: X-AIOS-Key.

Works with any AI tool. No plugins or extensions needed.
4

Start Talking to Your Data

AIOS auto-discovers every schema, builds a knowledge graph, and your agents can immediately query, update, and build databases in plain English.

Full data intelligence from the first conversation.
Connects to all major databases
SQL ServerSQL Server
PostgreSQLPostgreSQL
MySQLMySQL
MongoDBMongoDB
OracleOracle
SQLiteSQLite
SnowflakeSnowflake
BigQueryBigQuery
RedshiftRedshift
ClickHouseClickHouse
New — Vibe Coding for Data

Build Your Entire Database
in Plain English

Connect a database in 5 seconds. Then do everything — queries, updates, table creation, views, encryption, triggers, ETL — from any IDE or LLM you already use. No SQL required.

Simple

"Show me all unpaid invoices over $5K from the last 90 days"

"Create a Customers table with name, email, and signup date"

Intermediate

"Build a stored procedure that calculates monthly commission payouts by sales rep using the tiered rate schedule"

"Set up row-level security on Orders so each rep only sees their own"

Complex

"Normalize the legacy FlatOrders table into 3NF — split into Customers, Orders, OrderLines, and Products with FKs, migrate the data, and create a backward-compatible view"

Cross-Database

"Compare commission totals between QuickBooks and the CRM for Q1, flag discrepancies over $100, and write exceptions to a new reconciliation table"

DBA-Level Architecture

"Build a star schema data warehouse — create FactSales and dimension tables for Customer, Product, Time, and Region, write the ETL as stored procedures, and schedule nightly"

DDL, DML, encryption, partitioning, triggers, stored procedures, row-level security, indexing, normalization, ETL — every feature of all 10 supported database types, accessible in conversation. AIOS doesn't guess — it assembles context from 8 sources including your full schema, knowledge graph, sample data, and learned corrections to generate correct, database-specific SQL every time.

LangChain can run a SELECT. AIOS can architect your entire database in conversation.

Stop Managing Context Files by Hand

Every AI coding tool needs project context — and every one uses a different file format. Instead of maintaining static files scattered across every repo, AIOS stores all context in the database. Your IDE fetches it with one API call. Always current, never stale.

What tools expect
CLAUDE.md .cursorrules .windsurfrules AGENTS.md copilot-instructions.md .continue/rules/ .roo/rules/ .junie/guidelines.md .amazonq/rules/
Static files scattered across every repo
→
What AIOS provides
GET /api/aios/docs/bootstrap-{tenant}/raw
One API call. Stored in the database, not files.
Always current. Tenant-isolated. Feeds every tool.
  • Context lives in the AIOS database — not scattered across repo files
  • One API call returns the right context for any project, any tool
  • Schema knowledge, conventions, and corrections — always current, never stale
  • Corrections learned in one session apply to every future session across all tools
  • Context is scoped per tenant — no cross-contamination between organizations
  • Works with Claude Code, Cursor, Windsurf, Copilot, Codex, Cline/Roo, Continue, JetBrains AI, Amazon Q, and any tool that reads context files or HTTP
# GET /api/aios/ide/context/my-project

# My SaaS App

E-commerce platform with microservices architecture.

- Live: https://app.example.com
- Stack: Node.js, React, PostgreSQL, Redis

## Code Conventions

Use TypeScript strict mode. Prefer functional components.
API routes follow REST naming: /api/v2/[resource].

## Database Schema Context

- ecommerce contains Orders, Products, Users
- Orders.UserId joins to Users.Id
- analytics contains Events, Sessions

## Known Corrections

- Use dayjs not moment (deprecated)
- Auth middleware is at /lib/auth not /utils

---
Context generated by AIOS — always current

Meet the Brain

Most AI assistants answer from static knowledge. AIOS Brain answers from your live infrastructure — and can take action on it. It's the first conversational AI that operates your data estate, not just describes it.

  • Ask questions in plain English, get answers from your live databases
  • AI that manages AI — monitors your agents, reports what they learned, flags anomalies
  • "Email dusty a summary of today" triggers queries, composes, and sends — one sentence
  • Run SQL queries, traverse the knowledge graph, check schema changes — by asking
  • Morning briefing synthesizes all overnight activity into a narrative
  • Call it from your IDE, CLI, or dashboard — same API, full context
  • Two personas: Master Control Program or Kevin Flynn — your choice
# POST /api/aios/brain/chat
# Auth: X-AIOS-Key or Bearer token

> "What databases do I have?"

MCP: My sensors detect 41 databases across
the Grid. Finance: a5finance, figfinance.
CRM: crm-fig, bid-fig. ERP: erp-swansolutions.
Tools used: query_schema_sources

> "Email dusty a summary of today's activity"

MCP: Initiating data transfer. Grid analysis:
14 decisions logged, 3 corrections applied,
2 schema changes detected. Transmitting
summary to dusty@swank12.com... Complete.
Tools used: query_decisions, send_email

> "Show me overdue invoices from figfinance"

MCP: Executing query on figfinance...
Tools used: run_sql_query
| InvoiceId | Customer    | Amount   |
| INV-4421  | Meridian    | $120,000 |
| INV-4425  | Coastal     | $85,000  |

Seven Pillars of AIOS

Everything AI agents need to operate across your entire database estate.

Auto-Discovery

Point AIOS at any database — SQL Server, PostgreSQL, MySQL, MongoDB, Oracle, Snowflake, BigQuery, and more. It maps every table, column, primary key, and foreign key automatically.

Federated Queries

Query across any combination of databases in a single API call. Multi-step orchestration with access control per agent.

Semantic Search

Embeddings-powered search across all your schemas and knowledge. Ask "which tables store payment data?" and get precise answers.

Knowledge Graph

Auto-built graph of entity relationships from foreign keys. Traverse connections between databases, tables, and data flows.

Data Lineage

Track how data flows between systems. Know that when an invoice changes, it affects figures in three downstream databases.

Learning Loop

Agents log decisions. Humans correct mistakes. Corrections auto-embed into the vector store. Your AI gets smarter with every interaction.

AIOS Brain

A conversational AI that operates your data infrastructure. Ask questions, run queries, send emails, get briefings — from your dashboard or IDE. It manages your AI agents and explains your data in plain English.

Reusable Skill Packs

Catalog your Claude Code skills tenant-wide. Every .claude/skills/*.md capability pack becomes discoverable across projects — any agent can semantic-search existing skills before reinventing them. Push once, reuse everywhere, version-bumped automatically by content hash.

Where AIOS Fits

A single intelligent layer between your databases and your AI agents.

Your AI Agents

LangChain, CrewAI, OpenAI, Custom

↓

AIOS

Discovery • Graph • Search • Learning

↓

Your Databases

SQL Server • PostgreSQL • MySQL • MongoDB • Oracle • SQLite • Snowflake • BigQuery • Redshift • ClickHouse

Simple, Transparent Pricing

Start free. Scale as your agents grow. No credit card required.

Free

$0/mo
For vibe coders & side projects
  • ✓ 1 database
  • ✓ 2 agents
  • ✓ 5,000 API calls/mo
  • ✓ 5,000 embeddings
  • ✓ Community support

Builder

$29/mo
For serious builders
  • ✓ 3 databases
  • ✓ 5 agents
  • ✓ 25,000 API calls/mo
  • ✓ 25,000 embeddings
  • ✓ Email support
Most Popular

Pro

$99/mo
For teams & growing projects
  • ✓ 10 databases
  • ✓ 15 agents
  • ✓ 100,000 API calls/mo
  • ✓ 100,000 embeddings
  • ✓ Priority support

Business

$299/mo
For companies
  • ✓ 50 databases
  • ✓ Unlimited agents
  • ✓ 500,000 API calls/mo
  • ✓ 500,000 embeddings
  • ✓ Dedicated support

Enterprise

Custom
For large organizations
  • ✓ Unlimited databases
  • ✓ Unlimited agents
  • ✓ Self-hosted option
  • ✓ SSO / SAML
  • ✓ SLA + dedicated support

Ready to talk to your data?

Connect your databases in seconds. Manage everything in plain English from any IDE, LLM, or agent. Free to start.

Create Your Account →

Are you an investor?