End of Life for Inference, Fine-Tuning, and Dedicated Deployments
Legacy Content Notice
This article represents our previous focus on AI infrastructure and model hosting. We've since pivoted to focus on AI code review and verification - helping developers trust and improve AI generated and human written code.
As of July 24 at 11PM ET, we will be sunsetting our Inference, Fine-Tuning, and Dedicated Deployments services.
This wasn't an easy decision, but as this fast-moving ecosystem evolves, so must we. Going forward, we're excited to share that we'll be focusing all our energy on Guardrails for AI, starting with Verify, our new functionality for hallucination detection and improving LLM output reliability.
If you haven't tried it yet, Verify catches hallucinations, factual errors, RAG inconsistencies, and other inaccuracies before they reach users or trigger downstream actions. It's real-time, model-agnostic, and integrates easily into your existing stack (OpenAI, Claude, Mistral, LangChain, CrewAI, and more).
What this means for you:
-
Service Ends: July 24, 2025 at 11PM ET
-
Outstanding Balance: You can request a refund by sending an email to support@kluster.ai by July 31st, or leave it on-platform to use with Verify.
-
Fine-tuned models: You can request a copy of your LoRA weights by sending an email to support@kluster.ai by August 31st.
We're grateful for your support over the past year, and we'd love to help you transition to Verify. If you're interested, book a short call with our team. We're happy to walk you through it.