Lobu
NexusAi Summary: Lobu is an open-source control plane that gives AI agents verified company context via a shared memory graph and 60+ connectors. Build subagents, enforce permissions, automate workflows, and run behind ChatGPT, Claude, or Slack with event-sourced traceability.
NexusAi Overview
Connect Slack, Gmail, Jira, GitHub and more, then Lobu ingests threads, tickets, commits and contracts into a shared memory graph. Agents search that context, call connectors through MCP or SDK-built tools, and execute scoped actions. Automations schedule recurring checks while approvals and audit keep changes controlled.
Capabilities and architecture
Engineering leaders, operations teams, support and customer success use Lobu to ground agents in live contracts, tickets and release plans. Product and project managers orchestrate launch readiness checks across Slack, Jira and GitHub. Analysts build research subagents with constrained permissions, while security teams rely on approvals, scoped access and audit trails. If you already chat with ChatGPT or Claude, Lobu supplies the missing company context and tool execution without forcing users to change where they work.
- Unifies Slack, Gmail, Jira, GitHub and Notion into searchable, linked company memory.
- Runs subagents with pinned skills, prompts, permissions and scoped data access boundaries.
- Executes sandboxed code in Daytona-like environments with limited file and network permissions.
- Automates recurring research and QA checks using schedules, triggers and approval workflows.
- Captures event-sourced history, enabling provenance, audits and reproducible evaluations across iterations.
Why it matters
In practice
Start by connecting Slack, Gmail, Jira, GitHub or Notion using OAuth or API keys, then scope access per connection. Lobu begins syncing and indexing records into shared memory. Add MCP servers or build custom connectors with the SDK to expose domain tools. Create agents and subagents, pin prompts and permissions, and choose where they run: on device, in a research sandbox, or in the cloud. Pair the Mac app and Chrome extension for desktop and browser workflows. Configure automations, approvals and audit to match your organization’s policy.
NexusAi: Agents stop guessing because Lobu provides verified context, controlled tool access and a complete audit trail.
Setup and deployment
Many agent platforms bolt on tools; Lobu makes tools, memory and governance first-class. The shared context graph with append-only events provides durable provenance, while connectors and MCP bridge everyday systems without rewiring workflows. Subagents, sandboxed execution and approvals keep actions safe. Working behind ChatGPT or Slack minimizes behavior change, and self-hosting or cloud deployment meets different security postures. It is opinionated, inspectable, and ready for production iteration.
Open the tool and review its core product experience.
Create your account or access your existing workspace.
Use your own task to judge speed, quality, and fit.
Check similar AI tools before making a final decision.
Lobu
Lobu is built in the open under Apache-2.0 and focused on practical, company-aware agents. The team emphasizes connectors, permissions and observable runs over flashy demos, offering self-hosting or Lobu Cloud. Their goal: make agents useful at work by pairing shared context with controlled tool access.
Lobu is a shared-context platform for company-aware AI agents. It continuously syncs data from Slack, Gmail, Jira, GitHub, Notion and other sources into an inspectable memory graph, modeled as entities, relationships and append-only events. Agents and subagents query this graph, call connected tools through MCP servers or native connectors, and execute sandboxed code with scoped access. A Mac app and Chrome extension pair the agent with your desktop and browser, while the Connector SDK lets teams add custom tools. Deploy it self-hosted or on Lobu Cloud, run behind ChatGPT or Claude, and automate recurring work with schedules and triggers. Permissions, approvals and audit provide governance, and recorded runs plus corrections become reusable evaluation material for safe iteration. The result is a live, queryable company model that keeps context current across agents.
Visit Lobu
Comments (0)
No Comments Found