Level 01
Supervised
You review every action before it happens.
- Reads
- Asks
- Edits
- Asks
- Commands
- Asks
AI helpers for software developers
Meet Nx IDE.
Tell it what you need in plain English. Nx IDE plans the work with you, writes the code and checks it, right inside VS Code, a popular code editor. It works with the AI you choose, even one that runs on your own computers. By default, nothing changes without your approval.
31 ready-made AI specialists 26 AI providers, or run your own Sends us no usage data
Nx IDE installs into VS Code as an add-on called Nexgile Code, switched on with a Nexgile license key.
One platform
Nx·Studio designs your digital workers. Nx IDE helps your developers write software with AI. The Cognitive Engine makes routine decisions once it has proved it gets them right.
See how the platform fits together What it is
AI helpers inside your code editor.
You describe a task in plain English. Nx IDE reads your project, plans the work, writes the code and tests it. By default, it asks before changing anything.
The problem it solves
Fast help, without losing control.
Many AI coding tools leave the checking to you. In Nx IDE's Orchestrator mode, one AI specialist writes the change and two others test and review it. They are built to stop and ask before any work goes out.
Who it's for
Teams where every change must be checked.
Software teams at organizations with strict rules or high security needs, and teams updating older systems. They get AI help with approvals built in, controls an auditor can check, and no lock-in to one AI company.
How a task works
What happens when you give it a task.
Five steps, from your request to a checked result. You decide how much it may do on its own.
01
01
Describe the task
Type what you need in plain English. You can choose a specialist, for example Plan to plan the work or Debug to fix problems. Show it the files to read, such as PDFs and Word or Excel files.
02
02
Review the plan
For bigger jobs, the Architect specialist writes a plan first. Comment on any part and it rewrites just that part. Nothing is built until the plan is approved: by you, or after a countdown if you choose a more hands-off setting.
03
03
Approve, or let it run
By default, every change and command waits for your OK. A five-step setting called the Delegation Dial, and rules you write, can let some actions run on their own. Before saving, sharing or releasing code (committing, pushing or deploying), the AI is built to stop and ask. To block an action completely at every level, add a deny rule.
04
04
Every step is saved
It makes save points (checkpoints) of your files as it goes, so you can go back to any earlier step in one click. A live meter shows the estimated cost while it works.
05
05
It checks its own work
By default, when it has changed files, it reviews its own work, builds the project and runs your tests. Then it tells you what it changed.
your choice of AI
Use the AI you trust.
Pick from 26 AI providers, from the major AI companies to services that run open models, or run a model on your own computer. Switch whenever you like, with no lock-in.
Keep it all in-house
Run AI models on your own machines with tools like Ollama or LM Studio. Your code never has to leave your building.
Use what you already pay for
Use an AI chat subscription you already have, or the AI models that come with VS Code.
The right AI for each job
Each specialist can keep its own AI model, so planning, coding and reviewing each get the best fit.
Keys kept safe
The access keys for your AI accounts are kept in your computer's secure vault, not in settings files.
private by design
Nothing is sent to Nexgile.
Your code goes straight from your editor to the AI provider you picked, with nothing from Nexgile in the middle. Nx IDE sends us no usage data, and even the license check works offline.
No usage reports
Nx IDE sends no usage reports (telemetry). We never receive your code, what you type, or how you use it.
No middleman
Requests go directly to your chosen AI provider, or stay on your own computer if you run the model there.
An edition for US government buyers
A separate TAA edition, for the US Trade Agreements Act, leaves out restricted AI providers and models entirely. They cannot be switched back on.
Keep private files private
List any files or folders in a special “ignore” file (.nexgileignore), and the AI won't read them.
understands your code
It knows how your code fits together.
A feature called Cortex builds a private map of your code on your own computer. It shows how the parts connect and what a change might break, before you make it. It even covers older languages such as COBOL.
A map of your code
See what a change would affect before you make it. The map is built on your computer, offline.
Older code, too
It reads COBOL (an older language) next to modern code, putting old and new systems on one map.
Remembers your project
Your project notes carry over between sessions, and it can search what was agreed in past chats.
Reads what you read
It can read PDFs, Word and Excel files, and web pages. If you switch it on, it can also look at other code projects.
safety nets
Safety nets built into the software.
On top of approvals, Nx IDE refuses unsafe file changes and can limit where the AI may write. With the optional sandbox, your computer itself stops its commands from changing anything outside your project.
Save points
On by default. Save points (checkpoints) are made as the AI works, so you can go back in one click.
Clear boundaries
Rules can limit where the AI may make changes. Automatic tasks can be locked to certain folders, and the software enforces it.
A sealed-off sandbox
Turn it on for macOS or Linux, and commands stay inside your project and the folders you approve. You can also block their internet access.
No half-written files
It won't save a half-finished file. Changes that were cut off, or that skip code with a note like “rest of the file unchanged”, are refused.
specialists & skills
A specialist for every task.
31 built-in specialists, from Architect and Code to Tester, Reviewer and Shipper, each focused on one part of the job. You can add your own, or install more skills and tool connections from built-in marketplaces.
31 built-in specialists
Ten core roles plus 21 ready-made experts for everyday work, from tidying up code to changing database structures.
Make your own specialists
Describe a new specialist in a short text file. Your team reviews and tracks it like any other change.
Skills & tools in one click
One click adds a skill or a tool connection from the built-in marketplaces. Skills are instructions you can reuse, and you can write your own.
Follows your project's rules
Write down how your team works in a project file like AGENTS.md, and every specialist follows it.
work as a team
A whole AI team. Your final say.
In Orchestrator mode, one AI specialist (or a small team) writes the change, two others check it separately, and a person decides whether it goes out.
01 Code · or a team
Implement
One specialist, or a team of 2–5
02 Tester · Reviewer
Test + Review
Side by side, independently
03 You
Ship?
An approval question to you
04 Shipper
Ship
Saves it to a separate copy (branch)
Independent checks
The tester and reviewer start fresh, with only the request and the changed files, so the writer's reasoning can't sway them. Neither one can edit files.
Limited retries
If tests fail or changes are needed, a fresh specialist tries again, up to twice. Then the decision comes back to you.
One button stops them all
One Stop button halts every AI helper in the chain at once, and a map shows who is doing what.
Many AI helpers
Agent Manager
One screen to run several AI chats at once and watch each helper's progress, cost and questions.
Many AI helpers
Agent Teams
For larger jobs, 2–5 AI helpers (agents) split the work, for example the server side and the screens people see. They message each other and share a task list while you watch.
Many AI helpers
/batch
For a big job, it creates as many as thirty small tasks and works on several at once, each in a separate copy of your code. Any finished task can become a draft pull request (a proposed change) on GitHub.
Many AI helpers
/review
Up to five AI reviewers look over your changes at once, without changing anything. Fixes start only when you pick them.
The Shipper is instructed to save work to a separate branch and never overwrite anyone else's. “Ship?” is a question to you. At the Supervised and Trust Reads levels, it always waits for your answer. At higher levels, add a deny rule if shipping must always wait for a person.
Controls that are built in
The checks your team already uses, built in.
The approval steps your team already uses for code changes, built into the AI helpers themselves.
01
Approval
Human approval is on by default.
02
Written rules
Write rules that allow an action, ask first, or block it. Anything a rule allows shows which rule allowed it.
03
Asks before shipping
Before saving, sharing or releasing code, changing a database or deleting anything, the AI is built to stop and ask. A deny rule makes it a hard stop at every level.
04
Your own checks
Your own scripts (hooks) can run at set moments, for example to require your quality checks to pass before a task counts as done.
05
Protected settings
No AI helper can edit your permission, hook or plugin files. The main one asks before editing its own rule files.
06
Separate roles
The AI helpers that test and review code cannot edit files.
The software itself enforces tool access, rules, hooks and the sandbox. The AI isn't simply asked to obey them.
The security test
Three questions your security team will ask.
Who approved that change? What did it cost? Who was allowed to touch the code? Nx IDE has an answer for each.
Q1
Who approved that change?
A person did, or it ran under a setting or rule a person chose. Approval is on by default, and the AI is built to stop and ask before code is saved, shared or released.
Q2
What did it cost?
A live meter shows each task's estimated cost, and a Usage & Cost dashboard breaks spending down by day, AI model and project, with budget alerts. Optional limits pause the work until you decide.
Q3
Who was allowed to touch the code?
Only what your settings and rules allowed. Deny rules apply to every AI helper, including helpers started by another helper. None of them can edit your permission or hook files, and the tester and reviewer cannot edit files at all.
Most safeguards are settings or rule files. Your security team can read them, keep them with your code, and match each one to the controls they already use.
Visit the Trust Center the delegation dial
You decide how much it does on its own.
One dial under the chat box sets how much the AI may do without asking. Turn it up one level at a time as your trust grows.
Good to know: save points (checkpoints) are on by default at every level. Above Supervised, helpers started by another helper (such as the Orchestrator's specialists) run with every automatic approval switched on. Your deny rules still apply to them.
More review More autonomy
Level 02
Trust reads
Reading files inside your project runs on its own. Edits and commands still ask.
- Reads
- On its own
- Edits
- Asks
- Commands
- Asks
Level 03
Trust edits
Reading and editing files run on their own. Routine questions, including plan approval, get the AI's first suggestion after a 60-second countdown. Commands still ask.
- Reads
- On its own
- Edits
- On its own
- Commands
- Asks
Level 04
Trust execution
Reading, editing and commands run on their own, as do tools you have marked as always allowed. Starting more helpers, editing protected files or files outside the project, and a few other actions still ask. Deny rules still block.
- Reads
- On its own
- Edits
- On its own
- Commands
- On its own
Level 05
Autonomous
Everything runs without asking, with three exceptions. Your deny rules still block. Edits to protected files or files outside the project still ask, as do a few actions like fetching web pages.
- Reads
- On its own
- Edits
- On its own
- Commands
- On its own
A whole team in one tool
Meet the specialists.
There are ten main roles, and each one covers a different part of the work. For bigger tasks, the Orchestrator teams several of them up.
Architect
Plans and designs before any code is written.
Code
Writes, changes and tidies up your code.
Chat
Answers questions and explains how things work.
Debug
Finds and fixes problems, step by step.
Orchestrator
Runs a team: writer, tester, reviewer, shipper.
Tester
Checks a change independently.
Reviewer
Gives an independent code review.
Shipper
Saves and uploads approved work.
Spec Bootstrap
Interviews you and writes a project guide (wiki).
Spec Sync
Keeps the guide up to date as the code changes.
Plus 21 more for everyday tasks
RefactorTest GenDoc GenSecurityAPI ModDep UpgradeIssue FixerPR FixerMerge ResolverDocs ExtractorIssue InvestigatorIssue WriterDB MigratorDB ArchitectCode ArchitectDevOpsAPI TesterCode ReviewerTranslateUI DesignerDev
To add another, describe it in a short text file (Markdown) and it appears in the list, for one project or for all of them.
Everything it can do
Help at every stage of building software.
From the first plan to the final release. You decide how much it does on its own.
Plan & design
Plan the work before any code is written.
- A plan document you can comment on, part by part
- Sketch the overall design up front
- Build and keep up a project guide (wiki)
Build & update
Write, improve and update your code.
- Write, change and tidy up code
- Update older code, upgrade the libraries it uses, translate it
- Work with older code, including COBOL
Test & verify
Catch issues before your users do.
- Write tests for your changes
- A self-review and test run when a task ends
- An independent tester in every Orchestrator run
Review & release
Independent review, then a controlled release.
- /review — up to five reviewers at once
- Every edit shown for your approval, by default
- Asks before it ships, to a separate branch
Understands your code
Knows your whole project, not just one file.
- Cortex: a private map of how your code connects
- Memory between sessions, and search across past chats
- Search by meaning, after a one-time setup
Connects to your tools
Works with the systems your team already runs.
- Add tool connections to other systems
- One-click tools and skills from built-in marketplaces
- Optional: opens the app it built on your machine to check it works
Stays in control
Speed inside boundaries you define.
- Approval on by default; the Delegation Dial for the rest
- Written allow / ask / deny rules, and your own checks (hooks)
- An optional sandbox that walls commands in
- A Usage & Cost dashboard with budget alerts
Works at scale
From one chat to many AI helpers at once.
- Agent Manager: many chats side by side
- Agent Teams of 2–5 that coordinate
- /batch: up to 30 tasks, each in its own copy of the code
Works your way
Fits how your team already works.
- A save point at every step — restore in a click
- Optional typing suggestions and voice dictation
- Plain English → a ready-to-run command
- Chat in VS Code, or open it in a browser tab
Recently added
What's new in Nx IDE.
- Teamwork
Agent Teams, Agent Manager and /batch
Run many agents at once and let small teams coordinate. Split big changes into tasks that run side by side, each with its own draft pull request if you choose.
- Planning
Plans you can review, and /review
Review a written plan before work starts, then let up to five reviewers go over the code at the same time.
- Control
Written rules, hooks and a sandbox
Set which actions are allowed, need asking or are blocked. Add checks of your own at key points (hooks), and an optional sandbox on macOS or Linux to keep commands contained.
- Context
Memory, research and a cost dashboard
Notes that last between sessions, the ability to read web pages and documents, and budget alerts from the Usage & Cost dashboard.
- Models
Recent AI models
The latest models from major AI providers appear in the model list, ready to pick.
Good to know
What's on, what's optional.
Optional — off until you turn it on
- Operating-system sandbox (macOS and Linux)
- Per-task request and cost limits
- Web search and research in other code projects
- Opening the app it built in a browser to check it (needs Chrome or Edge)
- Automatic note-taking for project memory
- Typing suggestions, read-aloud and voice dictation (dictation needs the FFmpeg tool and a speech-to-text key)
Needs a one-time setup
- Searching your code by meaning
- Tool connections — add the ones your team uses, or install them from the marketplace
Will it act without my approval?
Not unless you let it.
Approval is on by default, and the Delegation Dial and your rules decide how much it may do alone. If you turn the dial up, helpers started by another helper also run without asking, within your deny rules.
Can it run up a bill?
You see the estimated cost in real time.
Daily and monthly budget alerts warn you but don't stop the work. Optional per-task limits pause it until you decide.
Which AI models can we use?
Your choice of 26 providers (21 in the TAA edition), including models that run on your own computers.
Your code never has to leave your building.
Does our code go to Nexgile?
No.
Nx IDE sends no usage reports (telemetry), and there is no Nexgile cloud in between. Your code goes only to the AI provider you choose.