Skip to content

The Next Generation of AI Intelligence

CHANAKAI is the layer on top of the AI model. Use it on your own device, across your company, or fully on your own servers for secure investigations.

Request a demoSee the three products

PARTNERED WITH

  • Sarvam AIPowered by Sarvam AI
  • Mistral AI
  • Anthropic
  • OpenAI
  • Microsoft Phi-3
  • Ollama
  • Connects with the tools you already use.Gmail, Google Calendar, Google Chat, Slack
AnyOpen model you bring
100%Processed on device
50+Report templates ready to file
A person working on a laptop and phone with AI tools

For individuals

An AI harness for day-to-day task automation.

A team working together around a digital table

For companies

Central control of how AI is used across your organisation.

Rows of server racks in a data centre

For police and government

Everything runs on your own servers. Built for secure investigations.

We are the harness, not the model.

CHANAKAI does not compete with the AI model you choose. It is the layer on top that turns the model into working tools, and it has its own price on top of the model's cost.

Bring your own model

Use a model you already trust, or run one on your own device.

Runs on your terms

Works offline, on your own machine or servers.

Tamper-evident audit log

A signed record of what was done.

Connects and automates

Connects to your existing APIs, MCP servers and databases, and can automate desktop tasks like filling documents and generating reports.

See it in depth

The details that matter..

Sample data
AccountStatus
SBI ··4821Reconciled
HDFC ··0937Structuring flagged
ICICI ··5510Reconciled

9 deposits of ₹49,000 in 6 days

Money-trail tracing

Reconciles bank statements automatically

Every statement is checked (opening + credits − debits = closing) before anything else runs. Pass-through accounts, fan-in patterns, and structuring under ₹10L / ₹50K are flagged automatically, with every figure tied to a real row.

Cited document answers

Every answer points back to its page

Ask a question across PDF, Word, Excel, PowerPoint or email, and the answer cites the exact page it came from. Click the citation, and that page opens.

When was the phone seized?
The phone was seized on 14 Mar 2026 at 06:40 PM from the accused's residence.Source: seizure-memo.pdf, p.4
Every Monday, pull new NCRP complaints and summarise them for me.
1Fetch new NCRP complaints
2Summarise each complaint
3Send the summary report
Runs weekly, Mon 9am

Automations & skills

Describe the workflow, not the steps

Write what you want done in plain words. CHANAKAI designs the steps, and runs it on a schedule : hourly, daily, weekly, monthly, or custom with a run log and retries.

Knowledge base search

Every document, indexed and searchable

Every file ingested across every case is indexed automatically, with entities extracted and visible, searchable by name, type, or status.

FileIndex
bank-statement-apr.pdfindexed
call-records.xlsxindexed
witness-statement.docxIndexing
EntitiesR. VermaHDFC ··0937+91 98xxx 21044
https://tools.example.in/mcp
Connected
search_recordsRun
fetch_documentRun
create_ticketRun

MCP tool connectors

Add external tool servers by URL; browse and run their tools directly.

  • Zoho Mail
  • Gmail
  • Slack
  • Cloudflare
  • Notion
  • Netlify
  • and many more…

Database Q&A

Connects PostgreSQL, MySQL, MariaDB or SQL Server. Plain-English answers, read-only.

  • PostgreSQL
  • MySQL
  • MariaDB
  • Microsoft SQL Server
  • and many more…
How many open tickets this month?
There are 38 open tickets this month, 11 of them opened in the last 7 days.
PostgreSQLRead-only
Summarise the witness statement.
The witness saw two people leave the branch at 04:15 PM and recognised one of them.
Ask anythingNo cloud call made

Local, private chat

Runs on locally loaded models, on the workstation. No cloud call to make.

Bring your own model

Add, test and switch models. Runs on CPU alone, faster with a GPU.

ModelsRuns on
Model AActiveGPU
Model BTestedCPU
Model CAddedCPU
TemplateCategory
Trace funds from a flagged accountMoney trail
Build a timeline from case filesTimeline
Draft an FIR from a complaintFIR drafting
Find the matching BNS sectionLegal section

Ready-made prompts

Templates for money-trail analysis, timeline extraction, FIR drafting and legal-section lookup.

For individuals

CHANAKAI Mini

An AI harness for day-to-day task automation. Bring your own model, or run one on your own device.

  • ₹199per month
  • Bring your own key, or an on-device small or large language model
  • Easy automation
  • Windows, Linux and Mac app
Try now

For companies

CHANAKAI Max

One place to control how AI is used across your organisation.

  • Centralised governance of AI
  • Implement automation workflows
  • Hyper-enable your tech stack
Learn More

Organisation scale

Built for large server deployments, at organisation scale.

Multi-user

Many people work on the same deployment at the same time.

Served on vLLM

Runs models on vLLM, a high-throughput serving engine that handles many users’ requests at once.

For police and government

CHANAKAI LEA

Everything runs on your own servers. No cloud at all. Built for secure investigations.

Case file interrogation

PDFs, scans, sheets and email answered in plain language. Every figure points back to its page.

This demo shows the LEA deployment. Mini and Max do not include this workflow. Shown here running on Sarvam AI. Works with the model of your choice.

CHANAKAICase 337
Ready · on device
Ask ChanakCase 337 · case file

Sample data. Nothing here is from a real case.

Custody is part of the product.

Custody logcase_337 · sample data
  1. 1Case 337 opened · 10:02 AMhash a41f…9c2e ← prev 0000…0000
  2. 2hdfc_4471_apr-sep.pdf added · 10:04 AMhash 7be0…13d4 ← prev a41f…9c2e
  3. 3Column mapping accepted · 10:05 AMhash c93a…70f1 ← prev 7be0…13d4
  4. 4Question asked in chat · 10:11 AMhash 2d6e…b8a0 ← prev c93a…70f1
  5. 5Report drafted (DOCX) · 10:19 AMhash f057…4e19 ← prev 2d6e…b8a0
  6. 6Report attested by officer · 10:22 AMhash 91cc…d2a7 ← prev f057…4e19
Chain verified · 6 entries, 6 signatures

Bring your own model.

Add any GGUF model from a file or a URL. CHANAKAI tells you whether your graphics card can run it before you commit. It ships set up with Gemma 4 12B and Qwen 3.5 9B. Works on CPU alone, and a GPU makes it faster.

Request a demo
  • Gemma 4 12BShips set up
  • Qwen 3.5 9BShips set up
  • Any GGUF modelFrom a file or a URL

CHANAKAI is available for personal use, enterprise scale, and government-grade secure deployment.

For individuals

CHANAKAI Mini

An AI harness for day-to-day task automation.
CHANAKAI Mini on a laptop, tablet and phone. The laptop shows the home screen: Good morning. How can I assist you today?, with tiles for Summarise documents, Extract key information, Draft a report and Analyse data.
  • ₹199 per month
  • Bring your own key, or an on-device small or large language model
  • Easy automation
  • Windows, Linux and Mac app
For companies

CHANAKAI Max

Central control of how AI is used across your organisation.
Dark server racks with a warm orange glow and a GPU server box in front.
  • Custom pricing
  • Centralised governance of AI
  • Implement automation workflows
  • Hyper-enable your tech stack
For police and government

CHANAKAI LEA

Everything runs on your own servers.Built for secure investigations.
A rugged black laptop showing a network map of connected points, with server racks behind it.
  • Custom pricing
  • Runs fully on your own servers
  • No cloud at all
  • Built for secure investigations
Works offline
Request a demo
Tamper-evident audit log
Request a demo
Works offlineTamper-evident audit log

Questions

About CHANAKAI

CHANAKAI is an AI harness. It is the layer on top of an AI model that turns the model into working tools. It is not a model itself.

Mini is for individuals. Max is for companies that want central control of AI use. LEA is for police and government, and runs fully on your own servers.

Yes. With Mini you bring your own key or run a model on your device. With LEA you bring any GGUF model, and it ships set up with Gemma 4 12B and Qwen 3.5 9B.

No. CHANAKAI is priced separately. A hosted model’s usage cost is separate.

About CHANAKAI LEA

No. Models, index and case data all sit on the workstation and inference runs locally, so there is no cloud call to make. The one optional network feature is sending a report by email, which uses your own department SMTP server on your own network and is queued locally if that server is unreachable.

A standard Windows workstation. It runs on CPU alone, and a GPU makes generation faster.

Not as it stands. The statutory templates have not yet been through legal vetting and ship behind a notice saying so. Everything the model produces is a draft an officer must verify and attest to. CHANAKAI never generates a signature or an attestation, by design.

A mapping is not accepted on the model’s confidence. It must pass arithmetic checks: the columns exist, dates and amounts parse, running balances reconcile, and where the bank prints its own totals, ours must equal them. If a check fails, the officer is shown which column is wrong and why.

See it run on your own machine

A live demo, with no cloud account and none of your data involved.

Which product are you interested in?
We reply by email. Your details are used only to arrange the demo.