Reviewed PRs in the client's repo inside one sprint. A system they keep when you leave.
DevIntern turns tickets in the client's existing tracker into reviewed pull requests in their existing repo, with a per-ticket record you can put in front of the steering committee. The pilot starts free on their machines, and at handoff the client keeps the config, the license, and the record.
Pilot timeline
their tracker · their repo
To start the pilot
interactive use is free
Team plan, in the client's name
up to 10 developers
Money-back on subscriptions
Team and Business plans
Why clients let it in
Everything runs on the client's infrastructure, with the client's credentials.
Most AI pilots die in procurement or the security review before they draft a line of code. DevIntern starts free and interactive on machines the client already owns, with credentials that never leave their perimeter, so the pilot starts this sprint instead of next quarter.
Nothing to procure to start
The pilot starts on the free interactive tier, on laptops and devboxes the client already owns. No new SaaS tenant to vet, no purchase order before the first pull request, no per-seat metering bolted onto the engagement.
Free tier · their hardware · no new vendor to vet.
Fits the stack they already run
Whatever tracker the client uses, DevIntern plugs into it: Jira, Linear, Trello, Asana, Azure DevOps, GitHub Issues, or plain markdown files. The coding agent is one config line, so it fits whatever AI contract the client already signed: Claude Code, Codex, Cursor, or OpenCode.
7 trackers · 4 agents · one config line to swap.
Credentials stay on client machines
Tracker tokens and repo access live in a project-local env file on the client's infrastructure, and each credential is sent only to its own vendor's API. Code, tickets, and diffs never touch our servers, and the source is available so their security team can verify that instead of trusting it.
Their tokens · their perimeter · their audit trail.
The worst case is "stops improving", never "stops working"
DevIntern is source-available under the FSL, and each release converts to Apache-2.0 after two years. Everything runs self-hosted on the client's machines, so the system you leave behind keeps running no matter what happens to us or to you.
FSL source-available · self-hosted · Apache-2.0 in two years.
The pilot playbook
Day 1 to sprint review: install, ship, hand off.
Both tools install locally in minutes and the interactive tier is free, so the pilot starts the day you do. Work real tickets in week one, and put the run history in front of the client before the sprint ends.
- Day 1 Checkpoint 1 / 3
Install & connect
- Install on a client laptop or shared devbox and wire it to the tracker and repo they already use. The interactive tier is free, so there is nothing to procure first.
- Run the first real ticket while their tech lead watches: feasibility check, implementation, self-review, draft PR in their repo.
- Nothing about their workflow changes: same reviewers, same CI, same tracker.
- Week 1 Checkpoint 2 / 3
Ship & enable the team
- Work real tickets end to end: the agent drafts, the client's engineers review, PRs merge in their repo.
- Show PMs @devintern/pm: Figma frames, logs, and rough prompts into codebase-grounded tickets in their tracker.
- Week 2+ Checkpoint 3 / 3
Report & hand off
- Present the run history at the steering meeting: tickets shipped, PRs merged, feasibility decisions, review outcomes. Evidence, not slideware.
- Move unattended automation onto a subscription in the client's name: Team covers up to 10 developers for $99 a month, less than a single engineer-hour, with a 30-day money-back guarantee. Scheduled pickup starts, and review comments start coming back as commits on the same branch, with replies.
- Hand off the config and operator playbook. The delivery record keeps accruing on their infrastructure after you leave.
The handoff
What the client keeps after you walk out.
A consulting engagement is judged on what is still working six months later. DevIntern's handoff is a set of concrete artifacts the client owns outright, not a dashboard license tied to your firm.
-
Run history on the client's disk
Every feasibility decision, draft PR, self-review pass, and review outcome, logged per ticket on their machines. This is the deliverable an auditor can open.
-
One-page operator playbook
How to queue tickets, rotate keys, pause a run, read the logs. Written so a tech lead owns it on day one without calling you.
-
Verifiable delivery record
Per-ticket evidence the client keeps after handoff: feasibility decision, PR link, review outcome. Verifiable artifacts, not slideware.
-
A license in the client's name
Interactive use stays free for their whole team. Unattended automation is a Team subscription in the client's name, $99 a month for up to 10 developers, with early pricing locked in for as long as they stay subscribed.
-
No vendor dependency on you
Nothing routes through your firm's infrastructure. If the client never calls you again, the system keeps running. That is the point.
Plugs into the stack the client already runs
No migration and no new pipelines: DevIntern reads tickets from whichever tracker the client already uses and opens pull requests in their existing repo. Review-comment automation runs on GitHub; Bitbucket PR creation is supported too.
Safety gates on every run
The guardrails the client's security team would have asked you to add are already there, on by default and documented in the handoff.
- Feasibility gate before any code: vague tickets get flagged back to the tracker with questions, not guessed at.
- Self-review loop on every diff: the agent reviews and fixes its own work before the client's engineers see it.
- With unattended automation on, review comments become commits on the same branch, with replies, until reviewers are satisfied.
- Crash and rate-limit recovery, with per-ticket run logs on the client's disk for audit and reporting.
The tools you deploy
Two terminal tools. One uninterrupted loop.
@devintern/pm Planning phase
Turns Figma frames, error logs, and rough prompts into codebase-grounded user stories posted straight into the client's tracker. Specs arrive clear enough to pass the feasibility gate instead of bouncing back with questions.
@devintern/code Execution phase
Reads a tracker ticket, runs the client's coding agent of choice (Claude Code, Codex, Cursor, or OpenCode) on the client's machine, and opens the draft PR. With unattended automation, it also answers review comments on the same branch. Swapping agents is one config line, so it fits whatever AI contract they already have.
Earn 20% on every referred purchase.
Track referrals, signups, and rewards from a personal dashboard built for advisors who recommend tools. Commission is paid on referred purchases, attributed at checkout: no attribution arguments, no reconciliation spreadsheets.
Pilot inside one sprint. Hand off a system the client owns.
Add DevIntern to your next engagement. Start free on the client's machines with the tracker and repo they already have, present the per-ticket run history at the next steering meeting, and transfer a license in the client's name at handoff. Subscriptions carry a 30-day money-back guarantee, so the downside of trying is a sprint, not a budget.
Self-hosted on client infra · their agent and keys · Team $99/mo for up to 10 developers · 30-day money-back.
Also evaluating DevIntern for another audience?