Jennefer
Jennefer IDEClosed beta in

Your AI dev team. In an editor.

Specialist agents plan, write, test and review your code on local models, in a desktop IDE. Nothing ships until you approve it.

Request access Read the docs

I turn your request into a plan before anyone writes a line.

See it running.

Real screens from Jennefer IDE, on one RTX 4070 with a local model.

Jennefer IDE: You approve every command. The Coder wants to run its test. Nothing runs until you say yes.
Closed beta starts in .Jennefer Hub and the IDE open first. Request access and be among the first in.Request access

Built for real engineering work.

A team, not a chatbot

Default roles plus library agents. Build and edit teams in Jennefer Hub > Agents, then pick one in the editor.

Plans before it codes

Each run goes from intent routing to a task plan, dispatching, verification and a closing report. A decision log records why.

You approve the risky parts

Approval cards ask before agents act; answer with Y, A or N. A workspace jail keeps agents inside your project.

Local models first

Pick a model and effort for each window, or set defaults in Jennefer Hub. The catalog shows what fits your GPU. Your own endpoint (OpenAI-compatible, Anthropic or Amazon Bedrock) is available on request.

A real editor around it

Monaco with live AI edits, diff view and preview, an integrated terminal, GPU and memory metrics, and a report of every changed file.

Plug in your tools

Connect local MCP servers so agents can use your tools, and let your own apps use Jennefer's AI through a local OpenAI-compatible API.

Eight roles. One crew.

Specialists in concert, in one workspace. Each role hands its work to the next and nothing moves on until the gate agrees.

Supervisor
Autonomous orchestration

Task decomposition, dependency scheduling and agent consensus.

Led by the whole crew
ArchitectLoop
Requirements and ADRs

Translates high-level prompts into deterministic tech specs.

Loop on duty
DeveloperByte
AST refactoring and codegen

Context-aware implementation and AST code mutations.

Byte on duty
QA TesterPatch
Test suite synthesis

Generates edge cases, fuzzing and unit test matrices.

Patch on duty
ReviewerPatch
Consensus gate

Enforces code standards, security audits and PR diff approval.

Patch on duty
DevOpsByte
Sandboxed execution

Local runtime isolation, dependency builds and migrations.

Byte on duty
UI DesignerPixel
Component styling

Tailwind tokens, CSS variables and design system sync.

Pixel on duty
DispatcherLoop
DAG execution

Directs parallel execution threads across available models.

Loop on duty

Every specialist has a face.

Meet Pixel, Loop, Byte, and Patch, the Nefers. They plan, build, and review inside Jennefer, and they have opinions about all of it.

  • Pixel UX Designer
  • Loop Planner
  • Byte Coder
  • Patch Reviewer
Meet the Nefers

More from the crew.

Coming soon

Jennefer Cloud

Disposable machines for your agents.

Isolated cloud sandboxes where agents install, run and preview without touching your computer.

Coming soon

Jennefer Studio

Images, video and campaigns on your GPU.

Make images, clips with sound, marketing designs and 3D models. Read any brand from its website.

Coming soon

Jennefer Workshop

Describe an app. Get one that thinks.

One sentence becomes a small Next.js app that can use your own AI model.