Memory that scales
with your agents.

One atomic fact stored in your cloud memory: a single statement the agent has decided is worth remembering. A 20-message conversation typically produces 10 to 30 memories, depending on how much new information the agent extracts. The dashboard shows the count in real time.

One read call: GET /v1/memories, POST /v1/memories/search, GET /v1/context, GET /v1/facts, GET /v1/users, or equivalent SDK methods (korely.search(), korely.get_context(), korely.get_facts()). Write calls (POST /v1/memories, PATCH /v1/memories/{id}) count only against the memories quota, not queries. Reads are retrieval, not generation: no generative model runs on the read path, which is why read quotas are an order of magnitude more generous than write quotas. The monthly counter resets the 1st of every month at UTC midnight.

No. An agent is a namespace for one of YOUR agent applications: your support chatbot is one agent, your sales copilot is another. The end users those agents serve are unlimited on every tier. Each memory and query is scoped to an end user via user_id, so a single support-bot agent can remember thousands of distinct customers. A user_id is just a label. An app with 10 users and one with 100,000 making the same API calls pay the same. What counts against quotas is total volume: memories written and queries made.

Soft cap with a 10% buffer, then rate-limited (429 response). You get an email alert at 80% of any quota (memories, queries, or agents) so you can decide whether to upgrade or trim before you hit the wall. We never auto-upgrade and we never bill overage.

Yes, and it is a product rather than an exception. You get the same engine on your own machine: Docker Compose, your Postgres, your model provider account, and a dashboard served by the install itself. The four public tiers stay hosted; self-hosting is the Self-hosted plan below, and it starts with a call.

On the public tiers, in our EU cloud, with data residency and DPA-covered sub-processors for the inference on the write path; you can erase everything about one user with one call. Self-hosted, it is your database on your machine and nothing reaches us at all. Extraction still calls a model, and you pick which and whose: Google, Mistral, or anything speaking the same API on your own network.

On-prem, the audit log and a DPA come with the Self-hosted plan. SSO works the way it does for anything you run yourself: your gateway authenticates your people and passes the result down, so your existing identity provider is the one that decides, not a second one of ours. What we do not offer is a standing SLA, because at our size that would be a promise about my nights rather than about the software; if you need one, we agree it in the call.

Same product, two surfaces. The consumer pricing at /pricing is for individuals using Korely as a desktop knowledge app (transcription, built-in AI chat, sync). The agents pricing here is for builders using Korely as the memory backend for their own agents via the REST API or SDK. Different tiers, different metrics, different audiences. You can subscribe to either, or both.