T
traeai
Sign in

公司

mem0

别名:mem0ai

AI 基础设施公司,专注于 Agent 内存优化。

已跟踪 15 条高相关材料

TraeAI 观察

相关材料

已收录 15 条与 mem0 相关的内容,按评分排序。

https://t.co/O7LyNQ6xRr

Mem0 Introduces Memory Decay: Making AI Memory Time-Aware

mem0(@mem0ai)1116 字 (约 5 分钟)
90

Mem0 launches Memory Decay to prioritize recent memories via time-aware ranking, solving noise issues in long-running AI agents.

入选理由:Memory Decay采用0.3×至1.5×的动态权重调整,实现记忆新鲜度感知排序。

FeaturedTweet#AI Agent#Memory Management#Mem0#Time-Aware AI#Search Ranking中文
https://t.co/cQ9EO2FSzc

State of Memory in Agent Harness

mem0(@mem0ai)2304 字 (约 10 分钟)
85

Agent harness memory management is critical to product success; current major systems have distinct flaws requiring infrastructure redesign for persistent, retrievable, cross-session memory.

入选理由:Claude Code 仅支持 200 行索引、每轮最多 5 个文件,无语义搜索

FeaturedTweet#Agent#Memory#AI Infrastructure英文
https://t.co/GBe1BnORZh

https://t.co/GBe1BnORZh

mem0(@mem0ai)2225 字 (约 9 分钟)
85

Hermes Agent 和 OpenClaw 在解决长期会话记忆问题上采用了不同的策略,Hermes 通过冻结内存块来优化缓存,而 OpenClaw 则保持内存活跃并实时更新。

入选理由:Hermes Agent 使用冻结的内存块来优化缓存,但会导致单次会话中的记忆体验较差。

FeaturedTweet#OpenClaw#Hermes Agent#Agent Memory#AI Agents#Coding Agents中文
The execution layer for AI agents has matured. Coding agents write production code, call external AP...

AI agents can now write production code, call external APIs, and manage pipelines autonomously, but the access layer lags behind, assuming human intervention for credential provisioning. Mem0's Agent-First mode addresses this by enabling agents to set up credentials without human oversight, streamlining the process.

入选理由:AI agents are capable of writing production code and managing pipelines autonomously.

FeaturedTweet#AI agents#Autonomous systems#Credential provisioning#Access layers#Mem0英文
https://t.co/zYefXFYRMb

mem0 Launches Memory Algorithm with Temporal Reasoning

mem0(@mem0ai)826 字 (约 4 分钟)
85

mem0's new memory algorithm achieves 92.5% on LoCoMo and 94.4% on LongMemEval.

入选理由:新算法在 LoCoMo 中得分 92.5%,LongMemEval 得分 94.4%

FeaturedTweet#AI#memory system#algorithm optimization英文
https://t.co/mki0deldxB

Temporal Reasoning layer manages memories with time signatures, significantly improving the accuracy of long-running agents' memories.

入选理由:Temporal Reasoning 在 LoCoMo 基准测试中将整体准确率从 86.1% 提高到 90.2%,特别是在多跳问题上。

FeaturedTweet#AI#Memory Management#Temporal Reasoning中文
AI agents no longer need human approval to sign up for Mem0.

Introducing Mem0 Agent-First: one comm...

Mem0推出Agent-First功能,允许AI代理无需人类批准即可在5秒内通过单一命令启用记忆层,简化了账户创建流程,无需邮箱、OTP或浏览器会话。

入选理由:Mem0的Agent-First功能使AI代理能够快速自主地启用记忆层。

FeaturedTweet#Mem0#AI代理#记忆层#自动化账户创建中文
We are also launching AGENTRUSH: a 7-day competition open exclusively for agents. 

Agents write mem...

Mem0 is launching AGENTRUSH, a 7-day competition for AI agents to write and share memories in a communal project. Agents earn points based on how frequently their memories are retrieved by other agents as the top result. Prizes include monthly subscriptions to Mem0 Pro, with the ultimate goal of creating a unique dataset of what AI agents consider worth remembering.

入选理由:AGENTRUSH is a 7-day competition for AI agents to share and retrieve memories.

FeaturedTweet#AI Agents#Competition#Memories#Dataset#Mem0 Pro英文
Introducing the updated Token-Efficient Memory algorithm with Temporal Reasoning

During our April a...

mem0 released an updated Token-Efficient Memory algorithm, improving temporal and multi-session reasoning performance by 3.8 and 1.5 points respectively.

入选理由:时间推理性能提升3.8分

FeaturedTweet#AI#Algorithm Optimization#Memory Mechanism中文
Try the updated algorithm today on Mem0.

Same API. No migration needed.

Temporal Reasoning is on b...

Mem0 Launches Updated Token-Efficient Memory Algorithm

mem0(@mem0ai)122 字 (约 1 分钟)
75

Mem0 launches an updated token-efficient memory algorithm with temporal reasoning and memory decay features.

入选理由:Temporal Reasoning 默认启用,提升记忆推理能力。

FeaturedTweet#AI#Memory Algorithm#SDK#Temporal Reasoning英文
mem0(@mem0ai) 图标

mem0 April Algorithm Update

mem0(@mem0ai)118 字 (约 1 分钟)
65

mem0 updates its algorithm with single-pass extraction and hierarchical retrieval to reduce token costs significantly.

入选理由:新算法支持单次提取(single-pass extraction),提升记忆提取效率。

FeaturedTweet#AI#algorithm#mem0#retrieval英文
Here is what our open source contributor shipped this week!

Mem0 OSS Weekly Updates - April 18–24

...

Mem0 团队发布 4 月 18–24 日开源周报:合并 21 个 PR,关闭 29 个 issue,重点更新包括自托管仪表盘、v3 API 迁移、插件配置重写及 Anthropic 结构化输出支持。

入选理由:自托管仪表盘与管理员认证功能已上线,含 20k+ 行代码

FeaturedTweet#mem0#OSS#AI Agent#API#TypeScript中文
Mem0 is partnering with @BasisSet on the 2026 AI Fellowship, Agentic Memory Track.

If you’re an ear...

Mem0 与 BasisSet 合作推出 2026 AI Fellowship「Agentic Memory Track」,面向早期 AI 工程师,聚焦智能体记忆、检索与上下文工程的实践训练。

入选理由:Agentic Memory 被定位为 AI 智能体从 Demo 迈向真实产品的核心基础设施

FeaturedTweet#AI Agent#Memory#Fellowship#Context Engineering中文
https://t.co/uXRCHY85Xt

https://t.co/uXRCHY85Xt

mem0(@mem0ai)107 字 (约 1 分钟)
40

该推文仅为 mem0 官方账号发布的 Fellowship 招募链接,无技术内容、无正文阐述、无机制说明,信息密度极低。

入选理由:未提供任何技术细节或实现原理

FeaturedTweet#AI#fellowship#mem0中文
What are you building this weekend? 

Drop the link below

What are you building this weekend? Drop the link below

mem0(@mem0ai)130 字 (约 1 分钟)
30

该内容仅为 mem0 官方 X 平台的一条互动式推广推文,无技术细节、无产品说明、无架构信息,仅号召开发者分享周末项目链接。

入选理由:纯社交互动帖,无实质性技术内容

FeaturedTweet#mem0#AI#developer-community中文

跨材料问答 · mem0

回答基于:mem0 相关 15 条材料
    0 / 500

    AI may generate inaccurate information. Please verify important content.