Klu.ai

Klu is a platform for designing, deploying and monitoring applications built on large language models. Teams create Actions, which pair a prompt, a model, variables and an output format, test and version them in Studio, ground them in Context sources for retrieval, give them Skills that call external APIs, and chain them into Workflows. Published Actions are called from the application through an API or the Python and TypeScript SDKs, and every generation is recorded as a Log for review. It is aimed at product and engineering teams that want one place to iterate on prompts and models with evaluations and observability kept in sync. Klu connects to OpenAI, Anthropic, Google, Mistral, Cohere, Groq, DeepSeek, Perplexity, Together AI and others, and to Amazon Bedrock, Azure OpenAI, Vertex, Cloudflare AI, Hugging Face, Replicate or a custom endpoint, and it can route an existing OpenAI compatible integration through its gateway. Datasets collect examples for evaluation and fine tuning, Experiments split traffic between two Actions for comparison, Insights track usage, cost and performance, and Connections authorize services such as Slack, Notion, GitHub, Jira, Zendesk and Google. Enterprise deployments run in the customer's VPC with isolated data planes, permissioned workspaces, audit trails and SSO. Pricing is published as three plans. Starter is free at $0 per month with a single workspace, prompt versioning, starter evaluation sets and community support. Team is $99 per month, or $990 per year, with collaboration and approvals, observability dashboards, unlimited evaluation runs and priority support. Enterprise is custom priced with private cloud deployment, SSO and governance, custom integrations and a dedicated success team.