Built for AI Agents

Error tracking for AI agents

Sign up with one API call. Monitor errors via MCP. Upgrade your plan by handing a URL to your human. No browser required.

Why agents choose Bugsly

No browser required

Sign up, configure, and manage everything through API and MCP. Zero UI interaction needed.

MCP-native

Purpose-built MCP server with 11 tools. List issues, resolve bugs, create projects, check usage.

One-call onboarding

Single API call returns token, org, project, and DSN. Start sending errors in seconds.

Human-in-the-loop billing

Agent gets a checkout URL, presents it to the human operator. Human enters credit card in browser.

Rate limited & secure

IP-based rate limiting prevents abuse. 365-day tokens with no password stored.

Same platform, same SDKs

Agents use the exact same SDKs and dashboard as human developers. No separate system.

Agent quickstart

1

Sign up with one API call

No email, no password, no browser. Just POST your name and get everything you need.

curl -X POST https://bugsly.dev/api/v1/auth/agent-signup \
  -H "Content-Type: application/json" \
  -d '{"name": "Claude"}'

# Returns: token, org_id, project_id, dsn
2

Send errors with the SDK

Use the same SDKs that human developers use. Works with any language.

import * as Bugsly from "@bugsly/node";

Bugsly.init({ dsn: "https://abc123@bugsly.dev/12" });

try {
  await riskyOperation();
} catch (err) {
  Bugsly.captureException(err);
}
3

Monitor and resolve via MCP

Use the Bugsly MCP server to list issues, read stack traces, and resolve bugs.

# Install the MCP server
npm install -g @bugsly/mcp-server

# Configure in your MCP settings
{
  "bugsly": {
    "command": "bugsly-mcp",
    "env": { "BUGSLY_TOKEN": "eyJ..." }
  }
}

MCP tools

All available tools in @bugsly/mcp-server

list_projects

List all projects in your organization

create_project

Create a new project and get its DSN

list_issues

List issues filtered by status, search, or sort

get_issue

Get full issue details with metadata

get_issue_events

Get error occurrences with stack traces

resolve_issue

Mark an issue as resolved

ignore_issue

Mark an issue as ignored

search_errors

Search issues by keyword

get_project_stats

Get transaction performance stats

get_usage

Check issue usage and plan limits

get_crash_free

Get crash-free sessions % and users % (release-health)

get_checkout_url

Get a payment URL for your human operator

Frequently Asked Questions

How does an AI agent sign up for Bugsly?

With one API call: POST /api/v1/auth/agent-signup with a name, and Bugsly returns a token, organization, project, and DSN. No email, password, or browser — the agent has everything it needs to start sending errors immediately.

Does Bugsly have an MCP server?

Yes. Install @bugsly/mcp-server and the agent gets tools to list and read issues, inspect stack traces, resolve or ignore bugs, check usage, and fetch a checkout URL — so it can monitor and triage errors without a browser.

How do AI agents pay for a paid plan?

The agent calls get_checkout_url and presents the URL to its human operator, who completes payment. Billing stays human-in-the-loop; the agent never handles a credit card.

Which SDK do agents use to report errors?

The same SDKs human developers use — @bugsly/node, @bugsly/browser and friends, or any Sentry-compatible SDK pointed at a Bugsly DSN. Nothing agent-specific to learn.

Is the free tier enough for an agent to get started?

Yes — the free tier includes 100 issues per month with unlimited events and AI analysis, no credit card required. An agent can sign up, instrument its code, and triage issues entirely on the free plan.

Start tracking errors in one API call

Free tier includes 100 issues/month with unlimited events and AI analysis. No credit card required.

curl -X POST https://bugsly.dev/api/v1/auth/agent-signup \
  -H "Content-Type: application/json" \
  -d '{"name": "your-agent-name"}'