AI Infrastructure for the Agentic Economy

Own Your Agents' Memory

Stop building data silos. Surf is the first File-First memory infrastructure that stores AI context as portable SQLite files in your customers' own storage.

✓Scale to 1M users

✓Zero storage liability

✓GDPR/HIPAA compliant

Multi-Protocol Integration

REST API or MCP Server—choose your integration pattern

REST API

Direct HTTP endpoints for add, search, update, and delete operations. Perfect for synchronous agent workflows and custom integrations.

MCP Server

Model Context Protocol for Claude, AutoGen, and LLM frameworks. Expose memory as native tools that agents invoke autonomously.

One Storage, Many Clients

Whether REST or MCP, your agent's memory lives in your customer's cloud. No vendor lock-in. No data silos.

rest-api.sh

1# Initialize Surf with your storage
2curl -X POST http://localhost:8000/api/memory/init \
3  -H "Content-Type: application/json" \
4  -d '{
5    "storage": "google_drive",
6    "folder_id": "your-folder-id",
7    "api_key": "sk_surf_..."
8  }'
9
10# Add context to memory
11curl -X POST http://localhost:8000/api/memory/add \
12  -H "Authorization: Bearer sk_surf_..." \
13  -d '{
14    "data": "User preference: likes detailed reports",
15    "user_id": "user_123"
16  }'
17
18# Search across memory
19curl -X POST http://localhost:8000/api/memory/search \
20  -H "Authorization: Bearer sk_surf_..." \
21  -d '{
22    "query": "user preferences",
23    "top_k": 5
24  }'

INTEGRATIONS

Connect Any Storage Backend

Bring your own storage. Surf works seamlessly with your customers' existing cloud infrastructure. No vendor lock-in. Pure data sovereignty.

Google Drive

OneDrive

SharePoint

AWS S3

Dropbox

Box

Notion

GitHub

Supabase

Snowflake

Need a custom storage backend? We support direct S3-compatible APIs.

From Raw History to Observational Memory

We don't just store data—we distill it. Transform massive chat logs into high-density, cost-efficient memory that scales with your agents.

3x – 40x

Compression

Our Reflector technology distills massive chat logs and tool outputs into high-density "Sovereign Observations."

90%

Cache Hit Rate

Stable, append-only memory structures mean you stop paying full price for repeated context.

10x

Lower COGS

We slash your primary cloud costs by an order of magnitude, making scaling a financial reality.

See how Surf cuts infrastructure costs while improving agent performance.

Infrastructure Built for Scale

A complete rethinking of memory infrastructure for autonomous agents

Ephemeral Sandboxes

Local-First GraphRAG

SQLite-vec + Graph stored in the user's cloud. No central database. Full data portability and independence.

Data Sovereignty

100% GDPR/HIPAA compliant by design. Your customers own their memory. Zero data residency concerns.

Zero Storage Tax

Pay for compute, not for bytes. Fixed monthly rate regardless of how many agents or how much memory they store.

Simple, Transparent Pricing

No per-token fees. No storage surprises. Flat-rate compute.

Compute-Only

Pay once, store forever

$99/month

Up to 10 concurrent agents

Unlimited stored vectors

Unlimited context windows

BYOS (Google Drive / S3)

Community support

All sandboxes included

Scale

Enterprise infrastructure

Custompricing

Unlimited concurrent agents

Dedicated infrastructure

99.9% SLA guarantee

White-label options

Priority support

Custom integrations

Advanced security features

vs Competitors: Pinecone and Mem0 charge per 1M vectors stored. With 1M users each storing 10k vectors, that's $10k/month in storage fees alone. Surf: $99/month, forever.

Ready to own your agents' memory?

Join AI teams building the next generation of stateful autonomous agents. Deploy your first agent in minutes.

No credit card required. Full access to all features.