Cleveri Enterprise
Elastic engineering capacity, inside your boundary.
Run Cleveri on your own infrastructure, with your own AI models, under your own engineering controls. Share Engineering Lanes across teams, environments and business units, and see exactly what every Job delivered.
- Private deployment on your VMs
- Your models, your keys
- Annual licence by Engineering Lane
- Unlimited users
Private deployment
Cleveri runs on your machines, next to your software.
Production code depends on databases, queues, internal APIs, private packages and company-specific configuration. Private deployment brings execution to that context. Source code and Job execution stay inside your environment.
- 01Install
Cleveri is deployed onto virtual machines in your cloud account or data centre. [Confirm supported clouds and install method]
- 02Connect
Point Cleveri at your work tracker, repositories, build and test commands and the internal services your code depends on.
- 03Choose models
Use your own provider keys or route every model call through your existing AI gateway.
- 04Scale Lanes
Each licensed Engineering Lane runs one Job at a time on your compute. Add Lanes when demand rises.
Bring your own AI
Your models. Your keys. Your gateway.
Cleveri doesn't lock you into a model. Bring the providers your organisation has already approved, and keep the agreements, data terms and spend controls you've already negotiated.
- Bring your own key and pay your provider directly.
- Route model calls through your existing AI gateway.
- Change models as better ones arrive, without changing how your team works.
- AI usage is recorded per Job, so model cost is tied to delivered work.
Security
Autonomy inside boundaries you define.
Cleveri executes engineering work. Your organisation keeps the boundary, the credentials and every decision that matters.
- Inside your network
Jobs execute on your VMs, beside your systems, under your network policies.
- Scoped credentials
Credentials are scoped to the work being performed. [Confirm]
- Human approval gates
Your engineer approves the plan before execution. Your team approves the merge.
- Platform-run verification
The platform runs your build, tests and checks. The coding agent does not decide whether its own work passed.
- Spending limits
Every Job has a spending limit. At the limit it stops at Needs Attention and a person decides. [Confirm]
- Your code stays yours
Source code and execution remain in your environment. Model calls go only where you route them.
PNG · 16:10 · 2400×1500
Must show: a repository profile with structure, conventions, build and test commands, and the context reused by the next Job.
Knowledge building [Confirm capability]
Every Job starts where the last one left off.
Each Job begins with investigation. Cleveri keeps what it learns about your repositories, so later Jobs start with context instead of from zero.
- Repository structure, conventions and dependencies
- The build, test and check commands your verification uses
- Approved plans and review outcomes, kept with each Job
- Knowledge stays inside your deployment
Automation
From work item to pull request, without hand-offs.
Cleveri runs the mechanics of delivery automatically. People step in only at the decisions: approving the approach, accepting the result and merging.
- Pick up development-ready work from your tracker [Confirm trigger options]
- Queue approved Jobs across every available Lane
- Run verification automatically on every change
- Deliver branches and pull requests into your normal review flow
- Work item readyYour team
- Plan proposedCleveri
- ApprovedYour engineer
- Queued to a free LaneCleveri
- VerifiedThe platform
- Pull request openedCleveri
Control
Intent stays human. Execution becomes elastic.
- Approval before execution
Nothing is built until your engineer approves the engineering approach.
- Shared capacity pools
Allocate Engineering Lanes across teams, environments and business units, from one shared limit.
- Spend you set
Spending limits per Job, AI usage per Job, and no automatic extra charges.
- Your merge, your rules
Cleveri delivers pull requests. Your reviewers and QA decide what is accepted and merged.
Observability
See what every Job did, and what it cost.
Capacity is measured, not assumed. Engineering leaders see the work moving through Lanes; finance sees cost per accepted change.
PNG · 16:9 · 2880×1620
Must show: Lanes in use by team, the queue, Jobs at Needs Attention, and Jobs delivered this month.
- Every Job: plan, approval, execution history, verification results and pull request
- AI usage and execution time per Job
- Lanes busy, Jobs queued and Jobs at Needs Attention
- Outcome measures: QA acceptance, cycle time, Jobs and AI cost per accepted change
Deployment options
One product. Three execution boundaries.
Cloud, Private and Enterprise are the same product, and your configuration carries over wherever possible.
| Feature | Cloud plans | Cleveri Private | Enterprise |
|---|---|---|---|
| Where it runs | Cleveri Cloud | Your infrastructure | Your infrastructure, across environments |
| AI models | Your key or Cleveri models | Your keys or gateway | Your keys or gateway |
| Capacity | Up to 10 Lanes | Licensed Lanes | Shared pools across teams |
| Engineering Jobs | Included allowance + extra Jobs | Unlimited within licensed Lanes | Unlimited within licensed Lanes |
| Users | Up to unlimited | Unlimited | Unlimited |
| Commercials | Monthly or pay per Job | Annual licence by Lane | Custom annual agreement |
| Getting started | Try it now | Pilot, then onboarding | Pilot, then onboarding |
How to start
Prove it on your own backlog, in your own environment.
- 01Scope
One repository and a defined set of development-ready work.
- 02Deploy
Cleveri in your environment, connected to your systems and models.
- 03Run
Your Tech Lead approves plans; your QA process accepts results.
- 04Measure
A report on QA acceptance, cycle time, supervision, Jobs and AI cost per accepted change.
FAQ