Currai
Observability and testing platform designed for large language model applications.
What is Currai?
The service captures detailed traces of LLM generations including inputs, outputs, model parameters, and associated costs. Users can group interactions by sessions or end users to analyze multi-turn conversations and monitor usage patterns across different models. Additional capabilities include nested tracing for complex chains involving retrievers and tools as well as plan-based data retention options. Integration remains straightforward through drop-in libraries that align with common standards without requiring infrastructure management. Teams benefit from dashboards that aggregate cost, token, and latency metrics while allowing traffic splitting for prompt experiments scored on real production data.
Key features
What you can use Currai for
Trace LLM Calls End-to-End
Instrument Python or TypeScript apps to capture every prompt, completion, token count, and tool call in nested traces for full visibility into chains and retrievers.
Run Production A/B Tests and Evals
Split live traffic across prompt versions by weight, then score cost and quality metrics directly on production traces without additional infrastructure.
Group Multi-Turn Conversations
Attach session and user IDs to traces so multi-turn interactions can be viewed together and sliced by end user or tags in self-serve dashboards.
How to use Currai
- 1Install the Currai SDK in your Python or TypeScript project
- 2Initialize the client with your public and secret keys
- 3Wrap an LLM call inside a trace and generation span
- 4Pass session and user IDs to group multi-turn conversations
- 5Point existing Langfuse or OpenTelemetry clients at Currai if migrating
Currai pricing
Pricing model: Freemium. Plan details are indicative — check the site for current prices.
STARTER
- 50 MB included
- 3-day retention
- Drop-in Python & TypeScript SDKs
- Full traces, tokens & cost in one view
- Langfuse & OpenTelemetry compatible
PRO
Popular- 2 GB included
- 14-day retention
- Everything in Starter
- Run evals and A/B test prompt versions in production
- Sessions & users roll-ups
- Cost, token & latency dashboards
BUSINESS
- 4 GB included
- 30-day retention
- Everything in Pro
- Hosted ingestion, storage & dashboards
- Priority support
Editor's verdict
Pros
- +Easy 5-line instrumentation
- +Compatible with existing Langfuse code
- +Includes cost and token tracking
Cons
- –Retention limited to 3-30 days by plan
- –Billed on processed data volume
Our take: Currai is a solid coding & dev choice. It's valued for easy 5-line instrumentation and compatible with existing langfuse code. The main trade-off is retention limited to 3-30 days by plan. A good pick if you want capable AI without a high upfront cost.
Frequently asked questions
Currai is an observability platform that provides full traces, token usage, and cost tracking for LLM applications along with built-in A/B testing and evaluation tools.
Summary
Currai is a solid coding & dev choice. It's valued for easy 5-line instrumentation and compatible with existing langfuse code. The main trade-off is retention limited to 3-30 days by plan. A good pick if you want capable AI without a high upfront cost.
User reviews
Verified reviews from the community shape this tool's rating.
Loading reviews…
Currai alternatives
Similar coding & dev tools worth comparing.
Explore & compare Currai
Data-driven comparisons, alternatives and rankings — kept current by our agents.
Featured in