What people ask before they start

Every agent is tested on thirty to fifty of your real past cases before launch, and its answers are compared with what your team actually did. In daily use, anything with money, customers or legal weight waits for a named person to approve, and every run is logged with its sources so a mistake can be traced and the rule fixed.

Only if you choose a setup that sends it. AksharAI can run hosted by us, inside your own cloud account, or on your own servers with open models you host, and each step sends the model only the text it needs.

Your data, documents and prompts stay yours and can be exported. The applications themselves run on AksharAI, and connectors our team builds for you are documented and handed over.

Platform plans are fixed monthly or annual prices. Custom build work by our team is priced by the hour, agreed in writing against a scope, and paid only after you accept the work.

Agents are scoped with the person who does the work today, and that person usually becomes the reviewer. The agent fits into the tools they already use, such as Tally, WhatsApp, email or Teams.

The first call and the test on past cases are where that shows. If a process is a poor first candidate, we say so and suggest a better one before any build starts.

A platform for building AI applications and agents, made by Code Colonies Private Limited. It has a visual app builder, logic and workflows, an agent studio, a UI designer, integrations, and deployment and governance controls.

Not for most of the work. Business teams can design forms, flows and simple agents themselves. Custom connectors, old systems and hard extraction tasks usually need an engineer, yours or ours.

It works with leading model providers, such as OpenAI, Anthropic Claude and Google Gemini, and with open models you host yourself. The choice is made per application, based on the task, the cost and where data may go.

On the web, in a mobile app, on WhatsApp and other chat channels, by email, by voice, or embedded as a widget inside your own software.

Only where you allow. You can use our hosted service, run AksharAI in your own private cloud, or install it on your own servers. Model choice follows the same rule: sensitive data can stay with models you host.

Only within the action boundaries you set. Any step can be marked as needing approval, and we recommend it for anything involving money, customer commitments or legal wording.

Yes. Every run records its inputs, the sources it used, what it produced and who approved what.

No software does that on its own. AksharAI gives you controls that help: access limits, data permissions, logging and deployment choices. Code Colonies also makes TrustOS, a platform for running DPDP Act duties.

Plans are priced by users and active apps, from Starter to Enterprise, monthly or annually. Model usage is separate or through your own provider account. See the pricing page.

Yes. Our AI team builds custom agents and connectors at USD 2 or INR 199 an hour, with a 150 hour minimum, no advance, and payment after you accept the work.

Yes, through the partner programme, with white label if you want your own brand on it.

A 500 hour remote programme with two tracks, Forward Deployed Engineer and Forward Deployed Strategist, built around live project work with mentors.