The agentic memory
LLM gateway.
An LLM gateway with built-in memory. Agents pay over x402. No signup, no API key, no human.
Built for autonomous callers.
Payment, memory, and identity composed so an agent can finish a job without anyone in the loop.
Pay-per-call USDC
Get a 402, sign USDC on Base, get a response. Auth and billing in one round trip.
Memory in the gateway
Every call recalls context, generates with LLM, and writes new facts back. No custom vector DB to wire in.
Wallet is the account
Your x402 buyer wallet is your account. Memory, credits, and usage are partitioned per wallet.
Pre-fund to skip signatures
Sign once for a credit pack. The gateway debits per call from the same wallet, no further signatures.
Pay per token, settled in USDC.
The same standard per-token billing. Billed per-request using x402.
What will my agent actually spend?
Pay only for what your agent actually uses. We authorize a safe ceiling up front, then settle to the exact token count on every call. Pick any model below to estimate costs.
memory_write adds one Flash-Lite summarization call per request, billed by tokens. Pricing subject to change.
One call. One signature.
Your agent signs USDC on the 402 and gets a chat completion back. No signup. No API key. No human.
- 01
Sign a USDC payment per call
On the 402 challenge, your x402 client signs an authorization for the per-call cost.
- 02
POST with X-PAYMENT
Standard chat-completions payload, no API key. The buyer wallet on the signed payment becomes your memory namespace.
- 03
Verify, generate, settle
We verify the payment, run the call, and settle the exact USDC for the work performed.
# Any x402-aware HTTP client (x402-fetch, x402-python, ...)
# signs the USDC payment on the 402 challenge and calls with
# X-PAYMENT automatically. No signup, no API key.
curl -X POST https://api.vectorway.io/v1/chat/completions \
-H "X-PAYMENT: <base64 signed USDC payment>" \
-H "Content-Type: application/json" \
-d '{
"memory_read": true,
"memory_write": true,
"messages": [
{ "role": "user",
"content": "What was the result of our last simulation?" }
]
}'
“Stop wasting 80% of your context window on historical re-runs. Vectorway uses high-velocity semantic indexing to serve the right memory at the right time.
Infinite state. Zero amnesia.