About Zep AI
Zep is a context layer for AI agents. It ingests business data, documents and conversations, builds a Context Graph for each user, customer, account or domain, and returns only the facts a given task depends on. Facts are dated and updated as the source data changes, and each record keeps a link back to where it came from.
The platform is aimed at teams building production agents that need reliable, governed memory. Zep applies access and retention policies in the data layer, decides what context each agent may retrieve, and logs every request for audit. The vendor states that retrieval takes under 200 milliseconds regardless of graph size, and that returning targeted facts instead of whole records cuts token use compared with tool calling. At larger scale, the Context Lake manages many graphs as one governed system and runs on Konig, Zep's graph database service. A Memory MCP Server is available, and plans include custom entity and edge types, webhooks and analytics at higher tiers.
Self serve plans are credit based. A free tier gives 10,000 credits a month with limits. Flex costs $125 a month with 50,000 credits included and then $25 per 10,000 credits, with 600 requests per minute and five projects. Flex Plus costs $375 a month with 200,000 credits included and then $75 per 40,000 credits, with 1,000 requests per minute and ten projects. The pricing page also offers annual billing at a 17 percent saving. Enterprise plans are custom and add negotiated rates, an SLA, SOC 2 Type II and a HIPAA BAA, and deployment in the vendor cloud or the customer's own VPC.