Zep logo

Zep

Paid
-(0 Отзывы)
0 Сохранено

Enterprise-grade memory infrastructure for AI agents based on temporal context graphs

Веб-сайт:https://getzep.com/
Добавлено:13 июл. 2026 г.
Посетители в месяц:
115.6K
Социальные сети и электронная почта:
AI agent memorytemporal context graphsContext Lakeenterprise AIagent infrastructurepersistent memory
Посетите веб-сайт
Zep
Посетите официальный сайт

Информация об инструменте

Что такое Zep?

Zep delivers persistent, temporal memory for AI agents via context graphs and a Context Lake that ingests sources, tracks facts over time, preserves provenance, governs access, and serves token-efficient context at enterprise scale.

Основные функции

  • Temporal Context Graphs

    Builds graphs from any source (chat history, business data, user interactions) to track entities, facts, and how they change over time, with automatic invalidation of contradicted facts while retaining history.

  • Context Lake

    Manages millions of context graphs as a governed, unified system powered by Zep’s proprietary Context Graph Engine, applying a data-lake pattern to agent context.

  • Sub-200ms Retrieval

    Maintains retrieval latency under 200 milliseconds (p95) regardless of graph size or count, scaling from 10K to 100M graphs.

  • Governed at the Substrate

    Provides attribute-based access control, retention policies, and audit/API logs that apply across every graph, query, and layer.

  • Provenance Preserving

    Traces every fact in the graph back to the source episode that produced it, enabling full audit of answers.

  • Observations

    Analyzes graph structure to surface patterns, recurrences, and co-occurrences as Observations that give agents a global perspective beyond individual facts.

  • Built-in Observability and Analytics

    Includes analytics for graphs created, users created, episodes added, and retrieval activity.

  • Flexible Deployment Models

    Supports Cloud, Cloud + Your Own Keys, and Bring Your Own Cloud deployment options.

Варианты использования

  • #1

    Persistent AI Agent Memory

    Provides managed memory of users, business data, and work done so agents can deliver reliable, personalized behavior at scale.

  • #2

    Long-Running Agent Context

    Supports long-running agent memory with temporal awareness, including benchmarks on LoCoMo and LongMemEval.

  • #3

    Enterprise Agent Stack Memory Layer

    Serves as governed, scalable memory infrastructure for enterprise AI agent deployments with compliance and audit needs.

Лучше всего для:AI teamsEnterprises building AI agentsDevelopers of production AI agents

Тарифные планы

Enterprise pricing is offered. Deployment models include Cloud, Cloud + Your Own Keys, and Bring Your Own Cloud. Book a demo and sign-up options are available.

Альтернативы

Claude

Claude

AI assistant for reasoning, coding, writing, and data analysis

🇺🇸24.5%
947M
Чат-боты и LLMAI assistant+10
Grok

Grok

Frontier AI chatbot and multimodal model family for advanced reasoning, coding, voice, image and video generation, and real-time search.

🇺🇸29.1%
28M
Чат-боты и LLMAI chatbot+11
DeepSeek

DeepSeek

Open-source language models and API services for coding, reasoning, and general AI tasks

🇨🇳47.9%
373M
Чат-боты и LLMLarge language models+11
Writingmate

Writingmate

Writingmate – a unified AI workspace with chat, file analysis, image, video and web search across 600+ models for $20/month.

🇺🇸11.2%
128K
Чат-боты и LLMAI+11
ChatPlayground AI

ChatPlayground AI

Run prompts across 30+ AI models side‑by‑side

🇺🇸66.7%
155K
Чат-боты и LLMAI model comparison+8
Blue Prism

Blue Prism

Enterprise agentic automation platform for RPA, AI agents, workflow orchestration, and BPM

🇺🇸13.2%
140K
Чат-боты и LLMRPA+11
Writer

Writer

An enterprise AI platform that lets teams delegate work to AI agents that execute on‑brand, compliant tasks end‑to‑end.

🇺🇸41.3%
274K
Чат-боты и LLMAI agents+10
Landbot AI

Landbot AI

No-code chatbot platform for AI-powered and rule-based conversations

🇵🇹13.9%
260K
Чат-боты и LLMAI chatbots+11

Часто задаваемые вопросы

What is Zep?
Zep is enterprise-grade memory infrastructure for AI agents based on temporal context graphs with sub-200ms retrieval and governance. It delivers persistent, temporal memory via context graphs and a Context Lake.
How does Zep provide memory for AI agents?
Zep builds temporal context graphs from sources like chat history, business data, and user interactions to track entities and facts over time. It preserves provenance, governs access, and serves token-efficient context at scale.
What deployment options does Zep support?
Zep supports Cloud, Cloud + Your Own Keys, and Bring Your Own Cloud deployment models.
How fast is retrieval with Zep?
Zep maintains retrieval latency under 200 milliseconds (p95) regardless of graph size or count, scaling from 10K to 100M graphs.
Is Zep suitable for enterprise AI agent deployments?
Yes. Zep serves as governed, scalable memory infrastructure with attribute-based access control, retention policies, audit/API logs, provenance tracking, and SOC 2 compliance for enterprise use.

Zep Контакты и социальные сети

Посетите официальный сайт

Отзывы

-

Поделитесь своим опытом использования этого инструмента...

Информация об инструменте

Модель оплатыPaid
КатегорияЧат-боты и LLM
Теги
AI agent memorytemporal context graphsContext Lakeenterprise AIagent infrastructurepersistent memory
Языканглийский
Последнее обновление16 июл. 2026 г.
Заявить права на этот инструмент