Trust & security

Trust is designed in, not added on.

You are handing us your company, your customers and your data. Here is exactly how we protect them, how our agents are governed, and what you own at the end.

An engineer checking a laptop in a quiet server room
Three principles

Everything else follows from these.

01

A person is accountable.

Every stage, release and decision has a named lead who signs it off. Agents prepare the work; people decide.

02

Your data stays yours.

You own your data, code and IP. We never use it to train models, and we never share it beyond the work you hired us for.

03

Secure by default.

Encryption, least-privilege access and audit logs are switched on from the first sprint, not bolted on before launch.

Your data

How we handle what you give us.

Ownership

Your data, the code we write and the IP we create are yours, written into the contract.

No training on your data

Client data is never used to train or fine-tune shared models. Model providers are configured so they cannot retain or train on it either.

Encryption

Encrypted in transit and at rest, with keys managed in your cloud account where you prefer.

Access

Least-privilege access, single sign-on and multi-factor authentication. Agents get only the data a task needs, and access is reviewed regularly.

Residency

India regions by default for Indian clients. On-premise, government cloud and private cloud options for regulated work.

Retention & deletion

Retention periods agreed up front. At the end of an engagement your data is returned and deleted, with confirmation in writing.

Sub-processors

The list of third-party services we use for your work is shared before you sign, and you approve any change.

Responsible AI

How our agents are governed.

Agents give us speed. These rules keep that speed safe. See each agent's guardrails →

Human sign-off

No agent makes a decision that affects a customer, a citizen, money or production systems. A named person signs every one off.

Full audit trail

Every agent action is logged with the task, the data used, the output and who approved it.

Tested before release

AI features are evaluated for accuracy, safety and misuse before launch, and monitored after.

Fairness checks

Anything that scores, ranks or routes people is checked for bias across groups before it goes live.

Clear disclosure

Your users always know when they are talking to an AI, and can reach a person.

Right model for the job

We choose models per client and per task, including open-weight models you can host yourself.

Compliance

Built to the rules you work under.

We design every engagement to align with the regulations that apply to you, and we document how.

DPDP Act 2023

Consent, purpose limitation, data minimisation, breach notification and data-principal rights designed into the product.

CERT-In directions

Log retention and incident reporting within the required six-hour window.

RBI requirements

For banks, NBFCs and fintechs, we work within RBI's IT outsourcing and cyber-security expectations, including audit access and data localisation for payments.

GIGW & WCAG 2.1 AA

Accessibility built in for public-sector and consumer products.

Security reviews

Send us your security questionnaire. A lead walks your team through our controls, one by one.

Delivery terms

What enterprises and governments can count on.

Integration

Works with what you run.

Your identity provider, ERP, CRM, data warehouse and core systems, connected through documented APIs.

Service levels

Agreed before we start.

Uptime, response and resolution times are set per engagement and written into the contract.

Handover

No lock-in.

Full source code, infrastructure scripts, documentation and decision logs handed over, with training for your team.

Incidents

A named contact and a runbook.

Escalation paths, reporting timelines and post-incident reviews agreed up front.

Built on

Proven technology, chosen per client.

These are technologies we build with. Listing them does not imply a partnership with their makers.

Cloud & hosting
AWSMicrosoft AzureGoogle CloudGovernment cloudOn-premise
Apps
ReactNext.jsNode.jsFlutterSwiftKotlin
Data
PythonPostgreSQLData warehousesEvent streams
AI
Frontier modelsOpen-weight modelsRetrieval on your dataEvaluation suites

Have a security questionnaire?

Send it over. A lead will walk your team through our controls.