Turn data into shared, structured memory

Owletto gives all your agents the same durable graph: connectors, recall, and managed auth without leaking credentials to the runtime.

Build Memory

How it works

Turn scattered prompts, tools, and application data into a shared context layer your agents can use everywhere.

01

Model the member graph

Represent members, companies, projects, repos, posts, topics, and introductions as linked objects so the community can remember who is building what and why they should meet.

Entities
Selected node
Member
Entity: Sarah Chen
Member
Type
Member
Name
Sarah Chen
Role
Founder
Company
Relay Labs
Relationships
member Sarah Chenworks_atcompany Relay Labs
member Sarah Chenbuilding_projectproject Relay Labs platform
member Sarah Chenmaintains_reporepository eval-orchestrator
member Sarah Chenwrites_aboutpost Why agent memory needs structure
member Sarah Cheninterested_intopic Agent memory
member Sarah Chenmatches_withmember Priya Natarajan
02

Connect sources

Ingest GitHub, LinkedIn, newsletters, personal websites, and manual profile forms through MCP proxying, public feeds, and Connector SDK integrations.

Community source inputs

Member context comes from the places people already publish work, identity, and intent.

TypeSourceAdded context
GitHubCode and repo activityTrack maintained repositories, contribution patterns, and technical areas of focus from connected GitHub accounts.
LinkedInRole and company contextPull current title, company, and professional background to keep the member graph aligned with real-world changes.
Newsletter / blogPublic writing and interestsUse Substack, RSS, and personal blogs to capture what members are actively thinking and writing about.
Profile importMember-provided contextCollect explicit goals, interests, and who the member wants to meet through forms or manual imports.
03

Let members connect their data

Use MCP login and OAuth for connected accounts, support RSS and public-site ingestion for newsletters and blogs, and allow manual profile imports without exposing credentials to agents.

How members connect accounts

Members connect accounts through MCP auth flows and operators can supplement that with public feeds or imports.

AccessSystemHow it works
MCP loginConnected accountsUse MCP/OAuth login for sources like GitHub and LinkedIn without exposing raw credentials to agents.
Public feedsWebsites and newslettersPull RSS, Substack, and public website content directly when a source does not require a private login.
Manual importProfile setupLet members or operators fill in a profile form or upload a structured import for goals, tags, and intro preferences.
Agent boundaryScoped accessThe community agent works with structured member context and approved workflows, not raw account credentials.
04

Reuse context everywhere

The same member graph powers community concierge agents in Slack, internal dashboards, and MCP clients like OpenClaw, ChatGPT, and Claude.

Community agents and workflows

The same member graph can power discovery, concierge, and intro workflows wherever the community already operates.

Community concierge
Answers questions like who should meet this week and why.
Slack
Member search agent
Finds members by topic, project, or recent activity using the shared graph.
Install to OpenClaw
Intro drafting workflow
Prepares warm intro drafts for Slack or email and waits for approval before sending.
Connect from Claude
05

Keep it fresh

A scheduled watcher turns new launches, posts, project updates, and hiring signals into suggestions about which members might care and which warm introductions to draft next.

Freshness watcher

A scheduled watcher keeps this memory current as new source changes arrive.

Opportunity matcherEvery 12 hours
Monitor connected profiles, newsletters, websites, and member updates for new launches, posts, hiring signals, funding news, and project changes. Identify which members are likely to care, explain why, and queue approved intro or outreach drafts.
Extraction schema
{ signals:[{ type, source, related_topics[], interested_members[], reason, suggested_action }] }
Schema evolution
Started with profile refresh and topic extraction. After repeated runs, added interested_members and suggested_action so the watcher could recommend who should see a launch, who should meet, and which outreach draft to prepare.

Beats other memory systems on public benchmarks

Apples-to-apples comparison on public memory datasets. Same answerer (glm-5.1) and same questions.

LongMemEval (oracle-50)

Single-session knowledge retention.

SystemOverallAnswerRetrievalLatency
Lobu87.1%78.0%100.0%237ms
Supermemory69.1%56.0%96.6%702ms
Mem065.7%54.0%85.3%753ms

LoCoMo-50

Multi-session conversational memory.

SystemOverallAnswerRetrievalLatency
Lobu57.8%38.0%79.5%121ms
Mem041.5%28.0%66.9%606ms
Supermemory23.2%14.0%36.5%532ms

Start building shared memory

Model the right entities, connect your sources, and keep long-term context available across every agent workflow.

Lobu on GitHub