Hand it spreadsheets or documents and say what you need. It builds the report or the spreadsheet, and you ask for changes in plain words. Runs on your server.
Open source · release v0.43.0 from
The difference is where the result lives: in text you carry over every time, or in a file you refine together.
Here is the May summary from the five files:
CapkaIts own Linux for every chat: it runs code, goes online and hands back a file that opens in Excel or Word.
Files are stored with you. Which model to plug in is your call.
One server for everyone. Each person has their own chats and files, and nobody sees anyone else’s.
What was given, what was asked, what came back.
Pull this into one table. Show me the deals where the margin is way below normal, and the clients who stopped buying in May. At the end, briefly: what should I check.
AfterWork out what we agreed with the contractor and how it kept changing. Deadlines, cost, what they were supposed to deliver. Show me where the documents and the emails don’t line up.
AfterLook at what changed for our four competitors this week. New features, prices, big announcements. If nothing much happened, just say so.
AfterCapka runs on your server, and every chat gets a separate computer on it. That is where it opens the files, runs code and keeps working after you close the tab.
The model sees only what the task needs, and you choose the model. With a local one, the content never leaves your server. How to harden the isolation is in the security guide.
Start on your own and invite the team to the same server later. Models and keys are set up once; chats and files stay each person’s own.
This is what the owner sees. Roles and spend are here; other people’s conversations are not in the admin panel at all.
Describe the task once and pick a time. Daily, weekly or monthly, Capka does it on its own and leaves the finished file. Every run starts from a clean slate, so an old conversation never leaks into a new report.
The web app is where the work happens. Away from your desk, the same Capka is in Telegram: the same files, the same context. Not a notifications bot.
Your procedures and templates become skills Capka reuses. Access to your databases and services is yours to grant.
You need an ordinary Linux server with Docker: 2 GB of memory to start, 4 GB recommended, a few dollars a month. The installer sets up Docker, pulls Capka and starts it.
A separate computercurl -fsSL https://raw.githubusercontent.com/LyoSU/capka/master/install.sh | bash
Prefer to inspect scripts first? Read the installer on GitHub. On first open: create an admin, connect a model, start working.
You are installing v0.43.0, released
An ordinary VPS. You pay the host directly.
You pay the provider for the tokens used. Everyone’s spend is visible in the admin panel.
None. Open source, no per-seat subscription.
Almost everyone can make files now. The difference is what they run on, where, and for whom. Pick what matters to you and compare for yourself.
Pick at least one condition.
CapkaEvery row is based on the product’s own documentation, linked as “source”. As of October 2, 2026. Capka’s side of the deal: you run the server and pay your model provider directly.
Every chat runs in its own Linux sandbox with separate file storage. Containers run unprivileged, without access to the host filesystem, and each chat’s context and files are kept apart from other chats and users. For public instances or untrusted users you can harden it further with gVisor or rootless Docker — see the security guide.
Two different things, and they are worth separating. Before a marketplace bundle installs, Capka shows what it will reach — which hosts its connectors talk to, which commands run in the sandbox, which secrets it asks for — and an update that widens that access stays blocked until you accept it. That is informed consent, not a restriction: the real boundary is the sandbox itself, where anything installed runs unprivileged and without access to the host filesystem. Admin policies — allow, ask, deny — and the audit log cover what happens afterwards.
Yours. Plug in providers you already pay for — Anthropic, OpenRouter, any OpenAI-compatible API — or run local models through Ollama. Use one shared key per instance, or separate keys per user.
Uploads, working folders and results stay stored on your Capka instance. Content needed for a task may be sent to the model provider you configure; with a local model, it can remain entirely inside your infrastructure.
Two bills, both paid by you directly: the server, and the model provider whose keys you plug in. Capka takes no cut and adds no per-seat fee, so a quiet month costs close to nothing beyond the VPS. Heavy file work costs more because it burns more tokens — the admin panel shows spend per person, so you see it before it surprises you.