Developers

SDKs and APIs for every stack.

Submit execution traces in minutes, query the execution corpus, and run evaluations against benchmarks derived from real-world agent runs.

Submit your first trace.

One POST request. Your agent is registered, and your first execution trace is captured.

  • REST API with JSON over HTTPS
  • Python and TypeScript SDKs
  • Anonymized server-side redaction (PII stripped at ingest)
  • Idempotent, retry-safe ingestion

Get an API key: khalidkhan46572@gmail.com

# pip install stratscope — coming to PyPI. Until then, install from source:
# git clone https://github.com/KhalidKhan541/stratscope && pip install -e sdk/python

import os
import stratscope

execution = stratscope.start(
    api_key=os.environ["STRATSCOPE_API_KEY"],
    base_url=os.environ["STRATSCOPE_BASE_URL"],
    project_id=os.environ["STRATSCOPE_PROJECT_ID"],
    agent_id=os.environ["STRATSCOPE_AGENT_ID"],
    model="gpt-4o",
    provider="openai",
    metadata={"task": "process_refund"},
)

execution.event(
    event_type="tool_call",
    payload={"tool": "search_tickets", "query": "refund #1234"},
)

execution.finish(
    status="completed",
    latency_ms=840,
    cost_usd=0.0012,
    tokens_in=1240,
    tokens_out=212,
)
npm install @stratscope/sdk

import { StratScopeClient } from "@stratscope/sdk";

// Coming to npm — until then, install from source:
// git clone https://github.com/KhalidKhan541/stratscope && npm install ./sdk/typescript

const client = new StratScopeClient({
  apiKey: process.env.STRATSCOPE_API_KEY,
  baseUrl: process.env.STRATSCOPE_BASE_URL,
  projectId: process.env.STRATSCOPE_PROJECT_ID,
  agentId: process.env.STRATSCOPE_AGENT_ID,
});

const exec = await client.startExecution({
  model: "gpt-4o",
  provider: "openai",
});

exec.event({
  eventType: "tool_call",
  payload: { tool: "search_tickets", query: "refund #1234" },
});

await exec.finish({
  status: "completed",
  latencyMs: 840,
  costUsd: 0.0012,
  tokensIn: 1240,
  tokensOut: 212,
});
curl -X POST https://stratscope-api.khalidkhan.workers.dev/v1/ingest/executions \
  -H "Authorization: Bearer $STRATSCOPE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"project_id":"...","agent_id":"...","model":"...","trace_id":"..."}'

# PATCH the execution when it finishes:
curl -X PATCH https://stratscope-api.khalidkhan.workers.dev/v1/ingest/executions/<execution_id> \
  -H "Authorization: Bearer $STRATSCOPE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"status":"completed","latency_ms":840,"cost_usd":0.0012,"tokens_in":1240,"tokens_out":212}'

SDKs and tools.

Beta — from source

Python SDK

pip install stratscope — Trace capture, batching and retry handling for any Python agent framework. Coming to PyPI; until then, install from source.

Beta — from source

TypeScript SDK

npm install @stratscope/sdk — First-class support for Node.js and edge runtimes. Coming to npm; until then, install from source.

HTTP

REST API

JSON over HTTPS with idempotent ingestion and rich query support.

Events

Webhooks

Real-time delivery of trace events, benchmark runs and dataset releases.

Tools

CLI

stratscope trace / stratscope eval — Inspect traces and run evaluations from your terminal.

Coming soon

Control plane

Datasets, evals and access management in one dashboard.

API reference.

EndpointMethodDescription
/v1/tracesPOSTSubmit an execution trace
/v1/agentsPOSTRegister an agent and configure sharing
/v1/datasetsGETBrowse and sample the execution corpus
/v1/benchmarksGETList benchmarks and methodology
/v1/evaluationsPOSTRun an evaluation against a benchmark

Get API access.

Tell us about your agents and we'll set you up with a key and a starter dataset.