Rivera
AI Agents · Coding Assistants · RAG

Memory your AI agents
never lose.

Rivera is a semantic memory engine: exact, deterministic vector search with zero indexing delay, plus grounded answers over everything your agents have learned. Store a memory and it's searchable the same millisecond.

Get your free API key Read the docs
Search

Exact, not approximate

No ANN trees, no index build step. Every query is an exact cosine scan — the same query always returns the same results, immediately after every write.

Answers

Grounded RAG built in

One call to /answer retrieves the relevant memories and generates a cited answer. No pipeline to assemble, no extra vendor keys.

Tenancy

Isolated by design

Per-tenant namespaces and API keys, hashed at rest. Your agents' memory is yours — scoped, revocable, and never shared.

Simplicity

Three primitives

Store documents, search them, ask questions. A clean REST API with an interactive explorer — integrate in minutes from any language.

Quickstart

# 1. Create a namespace
curl -X POST https://api.wirtel.ca/namespaces \
  -H "x-api-key: rv_..." -H "Content-Type: application/json" \
  -d '{"namespace_name":"my-agent","type":"text"}'

# 2. Store a memory (searchable instantly)
curl -X POST https://api.wirtel.ca/namespaces/my-agent/documents \
  -H "x-api-key: rv_..." -H "Content-Type: application/json" \
  -d '{"documents":[{"id":"1","text":"User prefers pytest over unittest"}]}'

# 3. Ask a grounded question
curl -X POST https://api.wirtel.ca/answer \
  -H "x-api-key: rv_..." -H "Content-Type: application/json" \
  -d '{"query":"which test framework does the user like?","namespace":"my-agent"}'