## What is PromptLayer?

PromptLayer is an LLM observability and prompt management platform that accepts OpenTelemetry traces over OTLP/HTTP. It ingests GenAI-standard OTLP spans and converts them into structured trace views, request logs, and analytics dashboards for monitoring and debugging LLM workflows.

### Key Features of PromptLayer

- **[Trace Observability](https://docs.promptlayer.com/features/opentelemetry#opentelemetry)**: Ingests OTLP spans and renders them as LLM call traces with inputs, outputs, token usage, and model metadata — visible under the **Traces** section of the dashboard.
- **[Prompt Management](https://docs.promptlayer.com/features/prompt-registry)**: Version-controlled prompt registry that links deployed prompts to observed traces for side-by-side comparison and evaluation.
- **[Analytics & Evaluation](https://docs.promptlayer.com/why-promptlayer/analytics)**: Aggregated dashboards for latency, cost, and quality across LLM providers and prompt versions.

## Prerequisites

Before integrating PromptLayer with TrueFoundry, ensure you have:

1. **TrueFoundry Account**: Create a [TrueFoundry account](/content/register/index.html) and follow the instructions in our [Gateway Quick Start Guide](https://docs.truefoundry.com/gateway/quick-start).
2. **PromptLayer Account**: Sign up at [promptlayer.com](https://www.promptlayer.com/) and have access to your workspace to retrieve an API key.

## Integration Steps

### 1. Get Your PromptLayer API Key

1. Log into your PromptLayer workspace.
2. Click **Settings** in the bottom-left sidebar.
3. Select **PromptLayer API Keys** under the Workspace section.
4. Click **Generate New API Key** and copy the key.

Create separate keys for different environments (e.g. **prod** and **dev**) so you can rotate them independently without affecting all integrations.

### 2. Configure OTEL Export in TrueFoundry

1. Go to **AI Gateway** → **Settings** in the TrueFoundry dashboard.
2. Scroll down to the **OTEL Config** section and click the edit (✏️) button.

3. Enable the **Otel Traces Exporter Configuration** toggle and fill in:

| Field | Value |
| --- | --- |
| **Toggle** | Enabled |
| **Protocol** | HTTP Configuration |
| **Endpoint** | `https://api.promptlayer.com/v1/traces` |
| **Encoding** | Proto |
| **Header Key** | `X-API-KEY` |
| **Header Value** | `<your-promptlayer-api-key>` |

4. Click **Save** to apply the configuration.

PromptLayer only accepts OTLP traces via its `/v1/traces` endpoint. It does not expose an OTLP metrics ingestion endpoint, so the Metrics Exporter should be left disabled. Make sure the header key is entered as **X-API-KEY** in all caps — PromptLayer’s API enforces this exact casing.

### 3. Verify the Integration

1. Make a request through the TrueFoundry AI Gateway (via the Playground or any API call).
2. Log into your PromptLayer dashboard and click **Traces** in the left sidebar.
3. Confirm that spans from `tfy-llm-gateway` are appearing with the full span hierarchy — including model calls, MCP gateway spans, and HTTP calls.

PromptLayer ingests spans asynchronously. Allow up to 60 seconds after your first gateway request before checking the Traces view.

## Configuration Reference

| Configuration | Value |
| --- | --- |
| **Traces Endpoint** | `https://api.promptlayer.com/v1/traces` |
| **Metrics Endpoint** | Not supported by PromptLayer |
| **Protocol** | HTTP |
| **Encoding** | Proto |
| **Auth Header Key** | `X-API-KEY` |
| **Auth Header Value** | Your PromptLayer API key |