Agentic AI workflows
Build, review then repeat

Nexus connects to your GitHub or Azure DevOps repos, lets you assign issues to AI agents like teammates, and shows you every tool call in real time. A second agent reviews every PR for security issues by default and leaves its findings as a comment — so nothing reaches your team unflagged.

The problem

Your backlog isn't waiting

There's always more in the backlog than there's time to build. Every ticket carries triage, review, and context-switching on top of the code itself — and typical AI assistants only help with that last part.

The solution

Agents that do the whole job

Nexus agents understand your codebase, read the issue, write the code, open the PR, and wait for your review. You stay in control; they do the work.

Everything your team needs to ship with AI

Built for engineering teams and individual developers.

Agents on issues and work items

Assign a GitHub issue or Azure DevOps work item to an agent the same way you'd assign it to a teammate. The agent reads it, explores the codebase, and opens a PR.

Real-time run monitoring

Watch every tool call, file read, and decision as it happens.

Built-in security review

Every agent-opened PR gets an automatic security pass by default — checking for OWASP vulnerabilities, hardcoded secrets, and unsafe patterns before it reaches your team. Turn it off per workflow if you'd rather skip it: it's a review aid, not a guarantee, so give agent PRs the same look you'd give any teammate's.

Multi-provider model support

Bring your own API keys for Claude (Anthropic), GPT-5 (OpenAI), Gemini (Google), or GitHub Copilot. Mix and match per agent role — use the best model for each task.

Team workflows

Create agent roles, assign them across teammates, and manage access on multiple repos. Nexus is built for teams, with shared visibility and per-repo controls.

Up and running in minutes

No infrastructure. No custom models. Just connect and assign.

  1. Connect your repo

    Connect your GitHub or Azure DevOps account and select the repos you want agents to work on. Takes under a minute.

  2. Configure agents with roles and models

    Choose a model provider, set the agent's role (frontend, backend, security, etc.), and tune behaviour with a system prompt.

  3. Assign an issue and watch it ship

    Assign any issue or work item to an agent. Monitor the run in real time, review the PR, and merge when you're satisfied.

Pricing

Free during closed beta

Nexus is free to use while we're in closed beta — up to 15 runs a month, 3 connected repos, and 5 agent configurations. Paid plans with higher limits, unlimited repos, and team features are coming soon — join early and help shape what we build.

Start free beta

Questions? Get in touch — or go back to Prometix.

Frequently asked questions

Is my code secure?
Nexus agents work inside a dedicated, single-use container that's created for each run and destroyed afterward. Your code is cloned only into that isolated, ephemeral environment for the duration of the run — never onto shared storage or persisted after the run completes. All access is scoped through your repository provider's permissions (GitHub App / Azure DevOps OAuth), and you control which repos are accessible at any time.
Which AI models are supported?
Nexus supports four providers today: Claude (Anthropic), GPT-5 (OpenAI), Gemini (Google), and GitHub Copilot. Bring your own API key for each — keys are encrypted before they're stored and are never shared with anyone else.
How is this different from GitHub Copilot?
Copilot helps you write code in your editor, one suggestion at a time. Nexus works autonomously on GitHub issues — it reads the issue, explores the codebase, makes decisions, writes code, and opens a PR. You review the output, not every line as you type.
Can I run Nexus on private repositories?
Yes. Nexus requests only the permissions it needs to read issues and open pull requests — whether that's via the GitHub App or Azure DevOps OAuth. You control which repos are accessible and can revoke access at any time.
What happens after closed beta?
We'll introduce paid plans with higher usage limits and additional team features — and there's a lot more on the roadmap beyond that.