Now in open beta · v2.4
AETHER gives your AI perfect recall over everything your team knows — messages, docs, decisions and code — retrieved by meaning in under 50 milliseconds.
No card required · 10,000 free memories · SOC 2 Type II
Powering memory for teams at
// Capabilities
Retrieval, ranking, sync and security — the infrastructure you'd otherwise spend two quarters building, exposed behind a single API call.
Ask in plain language. AETHER returns the exact passage, message or decision — ranked by meaning, not keywords — with citations your model can trust.
Every note, thread and commit is linked into a graph your agents can traverse.
p99 retrieval
A purpose-built vector index keeps recall under a blink, even across billions of tokens.
Bring your own model. One memory layer serves them all, hosted or on-device.
Encrypted in transit and at rest, scoped per tenant. Your memory is yours alone.
Connect a source once. New content is embedded and searchable the moment it changes.
// How it works
01 · Connect
02 · Recall
03 · Compose
Point AETHER at Slack, Notion, GitHub, Drive or any REST endpoint. We chunk, embed and index everything — and keep it fresh with incremental sync.
Send a natural-language query. AETHER ranks candidates across every source by semantic relevance and returns the passages that actually answer it.
Feed the retrieved context straight into your model. Answers arrive cited, current and impossible to hallucinate around.
// Ship it this week
Start free, wire up your first source in minutes, and scale to billions of tokens without touching a vector database.