telemetry.dev
An observability platform for AI agents that maps agent runs to a Trace/Span model and supports prompt management, Evals evaluation and alerting.
Timeline (last 90 days)
telemetry.dev received 89 votes on Product Hunt, ranking 4th, and is categorized under Analytics, Developer Tools and Artificial Intelligence.
telemetry.dev maps each run of an AI agent to a Trace, with each model call and tool call being a Span within the Trace, attached with latency, token usage, cost and error information.
telemetry.dev supports OpenTelemetry exporter integration.
telemetry.dev provides TypeScript and Python SDKs.
telemetry.dev unifies Prompt management and the Evals evaluation system on top of Traces, so that after modifying a Prompt you can run it directly on a dataset and compare scores against production.
telemetry.dev provides built-in alerting covering three types of issues: errors, cost spikes and agent loops.
telemetry.dev is suitable for teams already running AI agents in production and facing issues such as troubleshooting, Prompt iteration, cost attribution, or wanting to adopt the OTel ecosystem; for projects in the prototype stage with very small call volumes, it is less necessary.
All Facts (7)
telemetry.dev received 89 votes on Product Hunt, ranking 4th, and is categorized under Analytics, Developer Tools and Artificial Intelligence.
50%Unverifiedtelemetry.dev maps each run of an AI agent to a Trace, with each model call and tool call being a Span within the Trace, attached with latency, token usage, cost and error information.
50%Unverifiedtelemetry.dev supports OpenTelemetry exporter integration.
50%Unverifiedtelemetry.dev provides TypeScript and Python SDKs.
50%Unverifiedtelemetry.dev unifies Prompt management and the Evals evaluation system on top of Traces, so that after modifying a Prompt you can run it directly on a dataset and compare scores against production.
50%Unverifiedtelemetry.dev provides built-in alerting covering three types of issues: errors, cost spikes and agent loops.
50%Unverifiedtelemetry.dev is suitable for teams already running AI agents in production and facing issues such as troubleshooting, Prompt iteration, cost attribution, or wanting to adopt the OTel ecosystem; for projects in the prototype stage with very small call volumes, it is less necessary.
50%