Encrypted client-side
Material that may hold sensitive business or personal context is unreadable to Nenjo by design. Everything in between simply stores and routes encrypted data for your convenience.
Zero-trustCollaborativeOpen source
The private, collaborative agentic workspace.
Build and run teams of agents on a fully managed platform — with everything encrypted on your devices first, so not even Nenjo can read it.
Workers run on your infrastructure. Nenjo Cloud coordinates them — required today.
Managed dashboardEncrypted client-sideRuns on your workers
runtimenenjo worker
livepackagesPackage graph
Versioned knowledge and runtime units with explicit dependencies.
nenjo/agents/nenji@0.1.0 depends_on: knowledge/core@^0.1.0 abilities/design_agent@^0.1.0 abilities/design_workflow@^0.1.0Zero-trust by architecture
The convenience of a hosted platform, without the data trust it usually demands.
Nenjo runs a hosted dashboard and backend, so you get the polish of a managed product. But every chat message and dashboard-authored resource — agent prompts, knowledge documents, agent responses — is encrypted on your devices first. The platform only ever stores and routes ciphertext it cannot read, so “managed” never means “hand it over.”
You own the trust boundary
Your devices and your own Nenjo workers are the only places your content exists in the clear. Nothing in between, including the platform, can read it.
How the security model worksHow agents ship and evolve
Versioned packages, updated live on workers you control.
Nenjo gives agentic systems a real package manager: first-class, versioned packages for structured knowledge and runtime configuration that move cleanly between projects, environments, and machines you own.
Knowledge packages
Turn documents and domain context into living concept graphs with metadata-first search, graph traversal, and relationships agents can actually use.
Runtime packages
Ship agents, routines, abilities, tools, manifests, and policy as reusable units that move cleanly between projects and environments.
Live updates
Hot-swap packages and manifests while workers keep running. Update behavior without restarts or dropped tasks all from your desktop or phone.
Open source plus hosted control
Use the runtime in your code. Operate it from Nenjo Cloud.
Nenjo’s modular runtime and package primitives are open source. Nenjo Cloud adds the managed dashboard for onboarding, package operations, observability, and live updates when convenience and team workflow matter.
Open-source Nenjo
Core Rust crates, worker runtime, package primitives, manifests, and SDK paths you can inspect, extend, and embed in your own systems.
View the codeHosted control plane
Nenjo Cloud is the managed product dashboard for onboarding, managing packages, coordinating workers, and evolving live systems from one shared control surface.
Enter dashboardHybrid architecture
Built on Rust. Designed for your infrastructure.
Embed Nenjo core crates inside your product, run standalone workers with nenjo run, or combine both. The hosted control plane coordinates package operations while zero-trust boundaries and encrypted context keep runtime control close to your apps and infrastructure. Explore the harness →
Semantic versions
Explicit dependencies
Provider agnostic
Transport agnostic
Zero trust
Embeddable SDK
Documents become graphs.
Knowledge packages organize docs into concepts, relationships, and metadata so agents retrieve context by meaning instead of guessing from flat chunks.
- Concept graph traversal
- Metadata-first search
- Fresh context after package updates
Behavior ships as code.
Runtime configuration packages keep agents, routines, tools, and policies portable. Version them, review them, and roll them forward like product infrastructure.
- Portable manifests
- Explicit dependencies
- Runtime policy enforcement
Built for teams
Design, operate, and control your agents together.
Create a workspace, invite your teammates, and build agentic workflows together — with shared knowledge, clear roles, and explicit permission boundaries instead of scattered chat histories and one-off prompts.
Shared workspaces
Shared knowledge
Clear roles
Permission boundaries
How it works
Three steps to a live agentic system.
01
Run your workers
Deploy nenjo run where your data, tools, and policies already live. Install the harness →
02
Chat with Nenji
Nenji is included in the first Nenjo package to help you rapidly bootstrap agents, routines, abilities, and context from a conversation. Onboarding guide →
03
Evolve without downtime
Update packages or manifests live while running agents and routines continue uninterrupted.
Built for every builder
One control plane. Multiple ways to ship.
Product developers
Embed agents directly in your app with the SDK, then reuse the same packages across users, projects, and environments.
Explore the SDK →Personal AI Assistants
Run workers for long-lived assistants, multi-step routines, and capability-scoped agent systems on infrastructure you control.
Start onboarding →Consultants and teams
Package repeatable client solutions once, adapt per deployment, and keep every install inspectable and maintainable on your workers.
Read the security model →Package once. Deploy anywhere. Keep control.
Start with the open-source runtime, explore the docs, or enter the hosted dashboard to onboard and manage your live agent systems.