Skip to content
preview · endpoints may change before launch

Connect with a few lines.

OpenTelemetry in, proven PRs out. Six steps, and the only code you have to write is your exporter setup.

  1. ( 01 )Create a project

    Start by signing up free. Once you are in, create a project for one AI app and describe what it should do. Jev checks every trace against that. You get an ingest token that starts with sut_ing_. It is shown once, so keep it in your secret manager.

    Add your PRD, or let Suture read the docs in your repo, and it drafts checks for each feature. You review them and apply the ones you want. No docs? The AI features prompt has your coding agent write them from your code.

  2. ( 02 )Send OpenTelemetry (any language)

    Point any OpenTelemetry OTLP/HTTP exporter at Suture with these four variables, then use your language’s standard exporter. <token> is the ingest token from step 1. Keep the %20: the header value is URL-encoded, and %20 is the space after Bearer.

    shell / .env
    OTEL_EXPORTER_OTLP_TRACES_ENDPOINT=https://app.suture.nexprove.com/api/otlp/v1/traces
    OTEL_EXPORTER_OTLP_TRACES_HEADERS=Authorization=Bearer%20<token>
    OTEL_EXPORTER_OTLP_TRACES_PROTOCOL=http/protobuf
    OTEL_SERVICE_NAME=<your-agent>

    The endpoint takes protobuf (http/protobuf) or JSON (http/json), gzipped or not, and reads the OpenTelemetry GenAI and OpenInference attribute names, so most LLM instrumentation libraries work as they are.

    python / tracing.py
    from opentelemetry import trace
    from opentelemetry.sdk.trace import TracerProvider
    from opentelemetry.sdk.trace.export import BatchSpanProcessor
    from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
    
    provider = TracerProvider()
    provider.add_span_processor(BatchSpanProcessor(OTLPSpanExporter()))  # reads OTEL_* env vars
    trace.set_tracer_provider(provider)
    typescript / instrumentation.ts
    import { NodeSDK } from '@opentelemetry/sdk-node';
    import { OTLPTraceExporter } from '@opentelemetry/exporter-trace-otlp-proto';
    
    // Reads OTEL_* env vars. Start before the rest of your app loads.
    new NodeSDK({ traceExporter: new OTLPTraceExporter() }).start();

    [soon]A two-line helper for TypeScript and Python, suture-trace, is coming soon: not on npm or PyPI yet. Preview it on the tracing page; everything here works without it.

    Already on Langfuse? Connect it in the app with your Langfuse project keys, no code. On another tracer? Keep it, and add Suture’s exporter next to it.

  3. ( 03 )Map traces to code

    A fix needs to know which file a prompt lives in. Tag the span or the code, let Suture match prompt text against your repo, or map it once by hand in the app.

    either one
    # On the span that calls the model
    span.set_attribute("suture.prompt_id", "orchestrator.system")
    
    // Or next to the prompt in your code
    // @suture prompt:orchestrator.system
  4. ( 04 )Install the GitHub App

    Install the Suture GitHub App on the repository, pick an autonomy level (the default is pr), and Suture opens one small, proven PR per Issue. Nothing merges without you unless you turn on auto-merge, and that only covers prompt-only fixes of 30 lines or fewer with checks green.

  5. ( 05 )Optional: Issues in your editor

    The MCP server gives Claude Code or Cursor ten tools, so your agent can take an Issue from diagnosis to an applied patch. Create a personal token (sut_mcp_) in the app under Settings › API tokens; it acts as you, with what your role allows.

    shell / claude code
    claude mcp add --transport http suture https://app.suture.nexprove.com/api/mcp \
      --header "Authorization: Bearer <personal-token>"
    • list_projectsthe projects your token can see
    • list_issuesIssues for a project, latest activity first
    • get_issueone Issue with its diagnosis and recent fix runs
    • get_fix_planroot cause, evidence and the files to change
    • get_failing_casesredacted failing traces, to reproduce and test
    • get_patchthe diff a fix run produced, and whether it was verified
    • trigger_fixstart a fix run (needs a token with write)
    • aska plain-language question about your app
    • what_changedpass rate, new Issues, PRs, reverts and drift for a window
    • get_traceone redacted trace with its verdict
  6. ( 06 )Optional: the REST API

    The same data for scripts, CI and dashboards at https://app.suture.nexprove.com/api/v1: projects, Issues, traces, verdicts, fix runs and their patches, usage and drift alerts. Owners and admins create a project token (sut_api_) under Settings › API tokens. It reads by default, and starting a fix run needs write.

    shell / curl
    curl -s "https://app.suture.nexprove.com/api/v1/issues?status=open&limit=5" \
      -H "Authorization: Bearer <api-token>"
    curl -s "https://app.suture.nexprove.com/api/v1/fix_runs/fx_.../patch?format=diff" \
      -H "Authorization: Bearer <api-token>" | git apply --check

    Lists are newest first and page with limit and cursor. The OpenAPI 3.1 document is at /api/v1/openapi.json.