Future AGI: Observability, Prompts & Evals

Future AGI is a multimodal AI evaluation and observability platform that scores LLM outputs across text, image, audio, and video. Founded in 2024 in Mountain View, CA by Nikhil Pareek (CEO) and Charu Gupta, the company has raised $2.83M in funding including a $1.6M pre-seed led by Powerhouse Ventures and Snow Leopard Ventures with participation from 30+ angel investors.

The platform combines automated evaluation with production observability through several integrated modules: Evaluate provides proprietary accuracy metrics across modalities, Experiment enables no-code prompt prototyping, Monitor tracks real-time safety metrics for toxicity, bias, and policy violations, and Improve offers automated prompt refinement. Future AGI's TraceAI is an open-source tracing library built on OpenTelemetry that instruments 50+ AI frameworks including OpenAI, Anthropic, LangChain, LlamaIndex, CrewAI, and AWS Bedrock.

With a team of ~36 AI researchers and ML engineers from Microsoft and Amazon, Future AGI serves customers through both its SaaS platform and an AWS Marketplace listing. The platform holds a 4.8/5 rating on G2 with 12 verified reviews, with users particularly praising its multimodal evaluation capabilities and hallucination detection. The multimodal angle — evaluating image, audio, and video outputs alongside text — is a key differentiator that few competitors offer.