Last updated 07-24-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Respan (formerly Keywords AI)
Respan is an LLM engineering platform for teams shipping production AI agents and applications. Route model traffic through one gateway, trace every call in detail, run evals on offline datasets and live traffic, and monitor cost, latency, and errors from a single dashboard.
The product grew out of Keywords AI, a Y Combinator Winter 2024 company founded by engineers from the University of Illinois. What started as LLM routing expanded into full observability after customers asked for visibility into how agents behaved in production. Respan now positions itself around the loop from tracing to evaluation to iteration.
It fits ML engineers, platform teams, and AI product builders who need gateway failover, prompt version testing, production quality scoring, and spend controls without stitching together separate tools. Customers cited on the site include teams at Retell AI, Mem0, Lovable, and Gumloop.
One endpoint reaches 1,000+ models; switch providers by changing a single model name
Automatic fallbacks move to the next model when one errors or rate-limits
Response caching serves repeat prompts instantly and cuts cost on duplicates
Budgets and rate limits per API key, customer, or org with warn and block thresholds
Trace trees capture every LLM call, tool run, retrieval, and agent turn with cost and latency
Evaluators combine LLM judges, code checks, and human review into one weighted score
Alerts reach Slack, email, or webhooks when cost, errors, latency, or tokens cross limits
Free tier includes the full platform with 100k logs and no credit card required.
Single gateway endpoint for 500+ models with failover, caching, and spend limits built in.
Trace-linked evals let you compare prompt versions and score live production traffic.
SOC 2, GDPR, and HIPAA compliance with BAA available for healthcare teams.
Team plan starts at $199 per month when billed yearly.
Free tier caps logs at 100k with 7-day retention.
HIPAA compliance on Team requires a $249 per month add-on.
SAML SSO and self-hosted deployment are limited to Enterprise.
Does Respan have a free plan?
Yes. Respan offers a Free plan at $0 with no credit card required. It includes the full platform, 100k logs, 1k scores, 5 datasets, 2 evaluators, and 5 prompts. Sign up at platform.respan.ai.
What is the difference between Respan and Keywords AI?
Respan is the current brand for the same company, formerly known as Keywords AI. The product expanded from LLM routing into a full LLM engineering platform covering gateway routing, tracing, evals, prompt management, and monitoring.
How does Respan's AI gateway work?
Respan routes OpenAI-style calls through https://api.respan.ai/api/ to 500+ models, or supports provider passthrough endpoints. You can set fallback_models, enable response caching, and attach customer_identifier and metadata on every request. Each gateway call is logged automatically.
Can Respan score production LLM traffic automatically?
Yes. Respan runs evaluators on sampled production spans so quality regressions surface in live traffic, not only in offline tests. You can sample by percentage, filter by feature or customer, and set automations that alert when scores drop.
What integrations does Respan support?
Respan integrates with the OpenAI SDK, Vercel AI SDK, LangChain, LlamaIndex, LiteLLM, Mastra, Mem0, and many model providers including OpenAI, Anthropic, Groq, AWS Bedrock, and Google Gemini. Python and JavaScript/TypeScript SDKs are available.
Is Respan HIPAA compliant?
Respan is HIPAA compliant and offers a BAA for healthcare teams. On the Team plan, HIPAA compliance is available as a $249 per month add-on. Enterprise plans include custom packages and dedicated support.
How much does Respan's Team plan cost?
Respan's Team plan is $199 per month when billed yearly, marked as the popular tier for startups and growing teams. It includes unlimited datasets, evaluators, and prompts, a private Slack channel, and SOC 2 report access.