Claude partner in India, based in Pune. AI built directly on Claude.
woodfrog is an Anthropic Certified Services Partner, and a Claude partner in India with its own office in Pune. We design, test and run Claude agents for businesses across India, APAC, the UK and the US, with a person in control of every step that matters.

- Status
- Certified Services Partner, Claude Partner Network
- Built on
- Claude, by Anthropic
- Based in
- Pune, Maharashtra, India
- Delivering across
- India, APAC, the UK and the US
AI you can put in front of customers and auditors.
Five ways we put Claude to work.
Evaluations first. Then the agent.
- 01Define successMeasurable pass bars for each risk, agreed with you in writing.
- 02Build the evaluationsTest sets and graders, written before the agent is.
- 03Build the agentOn Claude, inside your systems, read-only until you approve.
- 04Run it with guardrailsApprovals, audit log, live monitoring and re-tests on every change.
What it changed for our clients.



About the partnership.
Is woodfrog a Claude partner in India?
Yes. woodfrog is a Certified Services Partner in Anthropic's Claude Partner Network, based in Pune, India. We build AI products and agents directly on Claude for clients in India and abroad.
How do I choose a Claude partner in India?
Ask for three things: a Claude Partner Network certification you can verify, Claude agents running in production with published results, and a written plan for how the AI will be tested before launch. We are happy to show all three on a first call.
Where is woodfrog based?
In Pune, Maharashtra, India, in our own office in Dhanori. Teams in Pune can visit by appointment, and we work with clients across India, APAC, the UK and the US.
What do you build on Claude?
AI agents that work inside your systems, voice agents in English, Hindi and Marathi, document and research assistants, and Claude on top of your own governed data.
Do you only work with Claude?
We build on Claude by default. For AI evaluation we test any model, including GPT, Gemini and open-source models, so you can compare them on your own tasks.