tools
Krutrim Cloud
Krutrim Cloud provides GPU and CPU compute, storage, networking, Kubernetes, AI Pods, and AI model services.
In inglese
Krutrim Cloud is a cloud platform for running applications, AI training and inference, agent workflows, storage, and data services. It offers CPU and GPU virtual machines, GPU bare metal, Kubernetes-managed AI Pods, object and block storage, networking, load balancing, and AI Studio.
Developers and businesses can use its web console, SDKs, APIs, CLI, and MCP integrations. Compute and most services are usage-based; storage, networking, load balancers, Kubernetes, and AI model calls can add separate charges.
Features
- Run CPU virtual machines for applications, APIs, backend services, and batch jobs
- Use NVIDIA A100 and H100 GPU virtual machines for training and inference
- Deploy Kubernetes-managed AI Pods with fractional or multi-GPU configurations
- Provision GPU bare-metal machines with direct hardware access
- Store data with object storage, block volumes, snapshots, and backups
- Manage VPCs, floating IPs, DNS, load balancers, and Kubernetes
- Access AI Studio for text, embedding, speech, multimodal, and evaluation workloads
- Use Python, Go, Java, Rust, CLI, MCP, and AWS-compatible APIs
Use cases
- Train and fine-tune machine-learning models on GPU instances
- Serve low-latency inference APIs from GPU or CPU infrastructure
- Run autonomous-agent code in isolated, disposable sandbox environments
- Deploy web applications, backend services, and batch jobs
- Store AI datasets, checkpoints, backups, and application files
- Run distributed AI workloads with Kubernetes-managed AI Pods
Pros
Cons
Capabilities
- Runs commands — “Create a sandbox, run a command, stream the output, and you’re done:” source
- VS Code — “Connect VS Code” source
- Command line — “Python MCP CLI Terraform Live terminal” source
- Choice of models — “Krutrim Cloud's AI Studio offers access to a diverse catalogue of open-source and in-house AI models for text generation, embeddings, speech, and multimodal use cases.” source
- API — “AWS-compatible APIs for zero-friction migration.” source
- Official SDKs — “Python, Go, Java, Rust, and MCP integrations.” source
- Runs models for you — “Run distributed training on GPU clusters, deploy low-latency inference, and fine-tune models with fully managed pipelines.” source
- Builds agents and workflows — “MCP server support, AI agent frameworks, and real-time streaming inference APIs.” source
- Traces and evaluates — “The cost of running Model Evaluations and Performance Evaluations on the Krutrim platform is the same as inference — based on the number of tokens processed.” source
Get it
Pricing
- Starting price
- ₹8/mo
- Prices checked
- 2026-09-25
Sandbox Compute — nano
- ₹1.56 /hr
- 0.25 vCPUs
- 0.5 GiB memory
- 20 GB disk
Sandbox Compute — small
- ₹2.91 /hr
- 0.5 vCPUs
- 1 GiB memory
- 20 GB disk
Sandbox Compute — medium
- ₹5.61 /hr
- 1 vCPU
- 2 GiB memory
- 20 GB disk
Sandbox Compute — large
- ₹11.01 /hr
- 2 vCPUs
- 4 GiB memory
- 20 GB disk
Sandbox Compute — x-Large
- ₹21.81 /hr
- 4 vCPUs
- 8 GiB memory
- 20 GB disk
Sandbox Volume storage
- ₹0.011 /GB-hour
- ₹8 /GB-month
GPU Instance — A100 80 GB × 1
- ₹189 /hr
- ₹148 /hr (Monthly)
- ₹132 /hr (6-month)
- ₹98 /hr (1-year)
- 96 GB RAM
- 80 GB GPU memory
- 24 vCPUs
GPU Instance — H100 × 1
- ₹213 /hr
- ₹198 /hr (Monthly)
- ₹186 /hr (6-month)
- ₹173 /hr (1-year)
- 200 GB RAM
- 80 GB GPU memory
- 24 vCPUs
GPU Instance — H100 × 2
- ₹426 /hr
- ₹396 /hr (Monthly)
- ₹372 /hr (6-month)
- ₹346 /hr (1-year)
- 400 GB RAM
- 160 GB GPU memory
- 48 vCPUs
GPU Instance — H100 × 4
- ₹852 /hr
- ₹792 /hr (Monthly)
- ₹744 /hr (6-month)
- ₹692 /hr (1-year)
- 800 GB RAM
- 320 GB GPU memory
- 96 vCPUs