Describe the outcome. The team ships it.
Jennefer hands each part of the task to the right specialist. They design, build and test in order, and a reviewer checks the result before it reaches you.
Pixel
Sketches the booking page
Jennefer is openWindows, macOS and Android
Your AI crew.
With your models.
Four AI specialists plan, design, code and review in your project, with the model you choose. Nothing ships until you approve it.
Pixel designs, Loop plans, Byte codes, Patch reviews. Click one.
Jennefer Hub
Watch the crew at work.
See the product before the details. 1:44 from brief to shipped.
No mockups. Real apps, running.
Every screen here is a capture of the app itself, on one RTX 4070 with Gemma 4 12B running locally. Cloud and Hangar show sample sandboxes and bots. Pick an app.
Jennefer Hub
Jennefer Hub
Available nowOpens to a chat with your model, and your apps open from its sidebar. In Settings you pick the model for each app, and the theme and text font for all of them.
More about Hub
The cloud sends a bill. Your GPU doesn't.
Jennefer agents use the model on your own graphics card, so there is no per-token charge. The Hub still counts every token and shows what a cloud model would have billed for the same work. Pick a Nefer's month.
110M input and 3.85M output tokens this month
Byte
That's a whole holiday. Staying local.
Estimated cloud bill
Claude Sonnet 5.5
Anthropic · Byte's month · 22 days
Prices as of
2026-10-19
Input tokens$220.00
110M × $2 per 1M
Output tokens$38.50
3.85M × $10 per 1M
Total a month$258.50
$3,102 a year. About 620 coffees.
On your GPU: $0 per token
Who sends the bill?
Same month, pick one- Your GPU$0 per token
An estimate, not a quote. Prices as of October 19, 2026 from each provider's public price page, at standard rates without cache, batch or off-peak discounts; hover a model for its source. Cloud models count tokens with their own tokenizer, so real numbers differ, and your computer still uses power. The months above are examples: Jennefer Hub > Usage shows your real token counts per app and model, and only the counts are kept, on your computer.
Hand over the goal. The crew does the rest.
Specialists plan, build and check the work on the model you pick, and you approve the result.
Stop counting tokens.
Local models have no per-token bill and no rate limit.
This week
Tokens0
Rate limitnone
Per-token feenone
Bill$0.00
Your models, your choice.
Run models on your device and switch per chat. Your own endpoint or API key is available on request, after a short review.
The boundary stays yours.
Local models run on your hardware. If you connect your own endpoint, its API key stays on your device. We don't train models on your code.
Your machine
No cloud model
Only what you need on your disk.
Local models and the Workshop pack download once, with pause and resume. Studio's images and videos run in the cloud, so they need no download.
Local modelInstalled
Workshop Packabout 480 MB
Jennefer Studio
Say it. See it. Ship it.
Images, video and whole campaigns with the best cloud models. Each one shows its cost before you make it.
Pick a prompt
FLUX 3 Klein, about 0.4 credits
Change it with words
FLUX.2 klein edits any picture. Drag the line to compare.
Make it move
Wan 2.2 makes quick clips. LTX‑2.3 adds sound, up to 20 seconds.
Fjellstad
Worth the early start.
Fjellstad
The autumn line is here.
Fjellstad
Made for cold mornings.
Ship the campaign
Paste your site once. Stories, posts and banners come out in your brand.
Studio Pro includes 115 credits of cloud AI a week; Free starts with credits, and All Apps Ultra goes much further.Prompts and images go to the model provider to make the picture; finished files are saved on your computer.
See plansJennefer Workshop
Describe an app. Get one that thinks.
One sentence becomes a small Next.js app. Apps ask your own model for answers, so no API key lives in the code.
Prompt
My apps / Focus timerWaiting
25:00
Session 2 of 4
Today
1 of 3 done
- Real code
- Next.js and Tailwind. Copy any app into a project and keep building it in the editor.
- AI built in
- Apps ask your model through the Jennefer bridge. Switch models and every app follows.
- Nothing to host
- Static apps with a small local store. No server code and no database to run.
One crew. Many apps.
Every part of Jennefer is its own app, made for one job. Hub ties them together: one account, one plan, and your Credits work in every app.
Available now for Windows and macOSJennefer Hub. Every app, one home.
Hub opens to a chat with your model, and you open your apps from its sidebar. It installs and updates them, and you sign in once. Download each model and pack once, and every app shares them without fighting over your GPU. In Settings you pick a model for each app, the theme (light, dark or system) and the text font for every app, and link your phone for Jennefer Mobile. Usage shows tokens and estimated cost.
IDE
Cloud
Studio
Workshop
Hangar- CLI
CLI+
Mobile
Jennefer Hub
Separate apps. One rhythm.
Each app does one job well, and all of them are built to work together from day one.
Send it anywhere
Make an image in Studio and drop it straight into a Cloud sandbox or an IDE project. No downloading and re-uploading.
One shared library
Images, videos, designs and apps you make land in one library that every Jennefer app can open.
Hand off between apps
When a Workshop app outgrows Workshop, move it into the IDE and keep building with the full team.
One account, one plan
Sign in once in Hub. Your plan and your Credits work in every app, and limits count once across all of them.
Ecosystem and security
In good company.
Startup programs from the companies behind the tools we build on: compute, sandboxes, identity, analytics and support.
Claude
Claude for Startups
Selected for the official startup program.
Atlassian
Atlassian for Startups
Selected for the official startup program.
MongoDB
MongoDB for Startups
Startup program partner.
Amazon Web Services
AWS Activate
Startup program partner.
Microsoft for Startups
Founders Hub
Azure runs the Jennefer backend and signs our Windows installers.
E2B
E2B for Startups
Runs the sandboxes in Jennefer Cloud.
Auth0 by Okta
Auth0 for Startups
Sign-in for Jennefer accounts.
PostHog
Scale Startup Cohort
Analytics on this website, only with your consent.
ElevenLabs
ElevenLabs for Startups
The Nefers' voices on this site, and voices and music for our product videos.
Sentry
Sentry for Startups
Startup program partner.
Notion
Notion for Startups
Startup program partner.
Deepgram
Deepgram for Startups
Startup program partner.
Mixpanel
Mixpanel for Startups
Usage analytics in the Jennefer apps, only if you turn it on.
BoldDesk
BoldDesk for Startups
Our support desk and the chat on this website.
Customer.io
Customer.io for Startups
Sends our account, billing and waitlist emails.
Datadog
Datadog for Startups
Uptime checks for the Jennefer backend.
Partner with us
Building developer infrastructure or tooling? Tell us how it could work with Jennefer.
Eight roles. One crew.
These are the roles in the IDE's agent team. Use the whole team or one member: the lead splits the work, the others build and check it, and you approve the changes.
Lead
Runs the team
Reads your request, splits it into parts and hands each one to a teammate. It reports back when the work is checked.
Led by the whole crew
Analyst
Writes the brief
Reads your code and notes first, then writes down what to build and how you will know it works.
Pixel on duty
Planner
Makes the to-do list
Turns the brief into small tasks with a checklist that fits your project, not a template.
Loop on duty
Developer
Writes the code
Edits the files, runs the code and tells you what changed and what is still open.
Byte on duty
Designer
Builds the screens
Builds screens in the style your project already has, and makes them work for every user and screen size.
Pixel on duty
DevOps
Runs the commands
Installs, builds and starts services. When a command fails, it reads the error before trying something else.
Byte on duty
QA
Runs the tests
Runs the checks that cover what changed, and tells a failing test apart from a broken setup.
Patch on duty
Reviewer
Reads the changes
Looks for bugs, security problems and regressions. By default it can only read, so it does not edit your code.
Patch on duty
In short.
A development environment where AI agents do real engineering work, on the models you choose.
No per-token bill
Agent teams on local models
Run agent teams on models on your own hardware, with the built-in local engine. There is no per-token rate and no surprise bill at the end of the month.
Your choice of model
Your models, your keys
Pick a local model in the built-in engine. Your own provider endpoint is available on request, after a short review. We don't train models on your code.
Jennefer Studio
Images, video and ads
Generate images, edit your photos, make clips with sound and design ads with cloud models like FLUX, Seedream, Kling and Veo. Nothing to download.
Agent teams
A team that checks its work
A lead splits the task between an analyst, a planner, developers, a designer and testers. Each one reports what it did, and a reviewer reads the changes before you do.
Coming soonLocal-first. No Jennefer servers. Just a license key.
Jennefer Premium is designed to run the agent team on your machine with local LLMs, activated with a license key and no Jennefer account. You can optionally add your own API key for a remote provider; prompts then go to that provider, never to Jennefer. Zero telemetry by default.
Not available yet. Features may change before release.
Jennefer Partners
Run a model cloud or publish open weights?
We add a short list of tested providers and local models to Jennefer. Apply to bring yours to every Jennefer app. There is no fee to apply.
Made by engineers who want a better way to build.
Jennefer started from one idea: development tools should give teams more control over their work, their machines and their ideas.
