Build AI agents on
your own data. On your device.
WebKaya is a Chrome extension for building AI agents grounded in your files — spreadsheets, PDFs, databases, web pages. Give an agent instructions, tools, and data; it works autonomously, shows its work, and never uploads a byte unless you choose. It combines the autonomy of an agentic assistant with a build-your-own-agent studio — executed entirely on the edge.
Runs with a built-in on-device model (Gemini Nano / WebLLM) out of the box · your own Ollama models · or bring your own cloud key.
One studio. Any agent. All on-device.
The same governed, audited engine — from patient records to threat logs to contract review — with every run brokered through an on-device blackboard and a downloadable proof.
Your best data is the data you can't upload
Cloud agent platforms need your data to travel to their servers, and every tool call burns tokens on someone else's meter. For patient records, deal pipelines, source logs, and anything under an upload ban, that's a non-starter. WebKaya flips it: the agent comes to your data, runs on your hardware at zero marginal cost, and proves nothing left.
Everything you need to build a trustworthy agent
A build-your-own-agent studio with the guardrails a cloud tool can't offer — because it all runs on the edge.
No-code agent builder
Name an agent, write its instructions, toggle its tools, attach its files. A canvas-ready workflow graph is generated behind the scenes — ready for multi-step teams later.
Grounded in your files
Drop in PDFs, manuals, and datasets. WebKaya chunks, embeds, and indexes them locally, so answers cite your sources — with nothing uploaded.
Verified by construction
Every run is brokered through an audited blackboard. Deterministic skills independently re-check numbers, so a weak model can plan but can't emit a wrong figure.
A local tool registry
Offline Python (Pyodide), verified analytics, local retrieval, and page reading today; out-of-core SQL and on-device vision landing next — each a toggle on your agent.
Multi-agent teams
Compose several agents into a phased pipeline — researcher → analyst → verifier — running concurrently and in isolation on the same blackboard.
Share as a .wkagent
Package an agent's prompts, tools, and data schema into one portable file. A teammate opens it and the agent recreates itself locally — data stays on each device.
Downloadable proof
Every run produces one audit document: which step read and wrote what, with provenance — verified, computed, model, or retrieved — on every value.
Zero marginal cost
Local models mean an agent can loop for hours without draining an API budget. Bring a cloud key only when you want frontier reasoning — masked first.
Local-only, enforced
Mark an agent local-only and it physically cannot make a network call — the runtime refuses to build a cloud provider for it. Not a promise; a property.
Ten agents you couldn't run in the cloud
Healthcare
Synthesize a patient's scanned records and CGM readings into one chart — under HIPAA, without uploading a byte.
Cybersecurity
Hunt anomalies in gigabytes of raw server logs locally — no network topology sent to a third-party model.
Finance & audit
Reconcile hundreds of receipt photos against a card ledger and flag every unmatched entry automatically.
Legal
Scan a 400-page merger agreement for non-standard clauses, grounded and cited — behind the firm's firewall.
Human resources
Rank hundreds of resumes against a role, indexed locally — no applicant PII leaves the recruiter's laptop.
Real estate
Blend millions of tax records with interior photos into a defensible listing-price recommendation.
Supply chain
Read messy vendor manifests and bin-pack the fleet for the fewest trips — right on the warehouse laptop.
Marketing & SEO
Crunch multi-gigabyte crawl exports to surface content gaps — without burning cloud API credits.
Food & beverage
Digitize a smeared clipboard inventory sheet and reconcile it against POS sales into a shrinkage report.
AI engineering
Stream a multi-GB eval log out-of-core and compute accuracy and failure breakdowns with local Python.
Build once. Run anywhere. Zero uploads.
1 · Build your agent
In the studio, give it a name, instructions, tools, and your files. WebKaya generates a governed workflow graph — no code required.
2 · Run it locally
Describe an outcome and step away. The agent plans, retrieves from your files, computes in a WASM sandbox, and independently re-checks the numbers.
3 · Trust the result
You get a dashboard, a grounded answer, or a report — plus a downloadable proof and a live meter showing 0 bytes ever left your device.
Not a promise — a property of the design
There are no WebKaya servers. Files are parsed in the browser, agents run on your own device, and a live meter proves the egress is zero. You can verify it yourself in DevTools.
- ✓ No servers, no analytics, no telemetry, no accounts
- ✓ Agents run on-device — built-in Gemini Nano / WebLLM, or your own local Ollama models
- ✓ Local-only agents physically cannot reach the network — enforced by the runtime
- ✓ Optional cloud models use your key and get locally-masked data with a visible amber banner
- ✓ Shared
.wkagentfiles carry the agent's shape, never your raw data
Files → local sandbox → your local AI. The internet path is severed.
Frequently asked
What exactly is WebKaya?
A Chrome extension and a platform for building AI agents that run entirely on your device. Ship-with-it agents (a verified Data Analyst and an offline Python Coder) sit alongside agents you build yourself — each with its own instructions, tools, and attached files.
Is my data really never uploaded?
Yes. WebKaya has no servers. Files are parsed in your browser, agents run on your device, and a live "0 bytes sent off-device" meter tracks it — verifiable in DevTools. An agent marked local-only can't make a network call at all: the runtime refuses to build a cloud provider for it.
How do I build an agent?
Open the builder, give the agent a name and instructions, toggle the tools it can use (Python, verified analytics, knowledge/RAG, page reader), and attach any files. WebKaya generates a governed workflow behind the scenes — no code needed.
What does "grounded in your files" mean?
Attach PDFs, docs, or datasets and WebKaya chunks, embeds, and indexes them locally. When you ask a question, the agent retrieves the most relevant passages and answers with citations to your sources — nothing uploaded.
Do I need an API key?
No. A built-in on-device model (Chrome's Gemini Nano where available, or WebLLM via WebGPU) works out of the box. You can point WebKaya at your own local Ollama models, or bring an Anthropic/OpenAI key for frontier reasoning — in which case data is masked locally first and a banner turns amber.
Can I share an agent with my team?
Yes. Export any agent as a .wkagent file — it carries the prompts, tool config, and the shape of the files it needs, never your raw data. A teammate opens it, re-attaches their own copy of the data, and the agent recreates itself locally.
Is it free?
WebKaya is free to install from the Chrome Web Store today. Paid Pro capabilities are on the roadmap.
Build your first agent in minutes
WebKaya is live on the Chrome Web Store. Install it, open the gallery, and run a built-in agent — or build your own on data you could never upload.
➕ Add to Chrome — freeChrome / Edge · Manifest V3 · No account needed · See the screenshots