Self-hosted · your machine, your model, your code

A coding agent working in front of you

A model writes your code, you decide what to keep. FastAPI behind, Vue 3 in front, one browser tab in between — and it puts that agent to work on a real checkout, in two pages.

Not · Just · Another · Editor

FastAPIVue 3Server-Sent Events MonacoMCPLSPPython ≥ 3.10Apache 2.0
127.0.0.1:8001
The WorkBench: the streaming conversation on the right, the agent's tool calls on the left, and the MCP and LSP tools it may use
The WorkBench — every tool call the agent made and the connector that answered, beside the conversation, streaming.

Two pages, one folder

The agent talks. You watch it land.

Both tabs are the same workspace. One drives the conversation, the other shows the tree the conversation is changing — so nothing the model does is hidden behind a diff you approve blind.

The WorkBench 💬

A streaming chat with the agent you pick. Chat has no tools at all; Code gets the file tools and the language servers; Architect reads your project and writes a spec, never your code.

An agent replaces the persona sent to the model, names the model the turn runs on, and decides whether a whole tool hub reaches it at all — and it can only ever narrow what Settings already allows.

agent modesbatch modestaged messages image attachskills catalogue
WorkBench · tool calls
A tool call expanded: the command the agent ran and the output the connector returned
Every call, its result and the connector that answered — readable, not hidden.

The Editor 🗂

This tab's folder as a file tree and a Monaco pane: create, rename, move and delete, save with a diff against the last commit, git status badges on the rows, and one terminal per connection on the platform's pseudo-terminal.

The code pane is served by the same language servers the WorkBench tools use — Ctrl+click a name for its definition, hover for its type, and real diagnostics in the status bar.

Monacoreal terminalLSP diagnostics diff vs HEADdated trash
Editor · same folder
The Editor: the same file tree, a Monaco pane and a real terminal in one screen
The Editor, same folder: the tree the agent is changing, and a terminal below it.

Everything streams

Answers, reasoning, tool calls and timings arrive over Server-Sent Events — the UI's only source of truth. When your endpoint answers ask again later, nojae waits out the pause and re-sends the identical request, so a provider outage resumes instead of ending a run.

What you get

Local by construction, not by promise

Your code and your endpoint stay on your machine. The agent's reach is bounded by a per-workspace read/write policy — and the connectors answer through it, so a folder you mark read-only is not editable by an agent either.

One command, no shell

The shell connector execs one command with no shell: chains, pipes and redirections are unrepresentable. Every other tool is read-only — asserted by a test, not by trust.

Batch mode

Works through a folder of task files unattended. A task is finished when its file leaves the lot folder — that move is the agent's own acknowledgement, and a file still there is relaunched up to the budget.

Nothing is destroyed

Every save keeps the bytes it replaced in a backup, and a delete moves the file into a dated batch under .nojae/trash/. A restore is one mv away.

MCP and language servers

Connectors and language servers are declared per workspace in a file, hot-toggled from the UI, and mirrored by two hubs that share one path policy with the Editor.

Skills

Drop a SKILL.md folder in .agents/skills and the agent starts every conversation knowing it exists — name and description only. The body is read on the model's own judgement.

Honest meters

Session clock, generation speed and context occupancy, measured app-side on chunk arrivals — so tool execution never enters the numbers, and a 40-minute outage is not a 40-minute session.

Up close

The parts you actually touch

Click any screenshot to enlarge it.

Quick start

Three commands, then one setting

You need Python ≥ 3.10 and Node.js. Desktop browsers only.

bash
# install
python -m venv .venv && source .venv/bin/activate
pip install -e ".[mcp,lsp]"

cd web && npm install && npm run build && cd ..

# run
nojae

# → open http://127.0.0.1:8001
  1. Install the extras you want

    mcp brings the WorkBench connectors, lsp the language-server tools, windows the terminal's pseudo-console driver. None is mandatory.

  2. Start the server

    One command, and it serves the built front end on 127.0.0.1:8001.

  3. Add one model

    Settings → Servers: your OpenAI-compatible endpoint — http://127.0.0.1:8080/v1 for a local llama.cpp — press Fetch models, add a model, give it to an agent.

A fresh install has no model, and no turn can run until you give it one. Nothing else is required for the Editor; the connectors only decide how far the agent may reach.

Read the code, run it locally

Apache 2.0, self-hosted, and the whole thing is one FastAPI app in front of one Vue 3 build.

« On ne dit kenavo qu'à ce qui peut nijañ. »