Entries describe capabilities available in the product today.
An agent can mark one issue resolved or ignored, reopen it and add a comment. Resolving or ignoring requires a note, which is stored on the issue. search_events accepts an issue id so the agent can see whether that issue is still firing. Setup pages cover Cursor, Claude Code, VS Code, Codex, Gemini CLI and Windsurf.
Dashboard and API responses, AI payloads, MCP results, exports and replay playback count against a data-transfer allowance. Developer includes 100 MB and hard-stops at 125 MB. Team, Business and Enterprise include 1 GB, 5 GB and 20 GB, then $0.10 per GB.
Docs cover the hosted MCP server at https://api.newtalaria.com/mcp. Sign in with OAuth or a personal token from Settings → Integrations. The agent reads errors, traces, sessions and project health, and can change safe project settings, mint a key and roll a flag.
Analytics → Heatmaps shows where visitors click and how far they scroll, over a masked snapshot of the page for desktop, tablet and mobile. Filter to rage, dead, error, first or last clicks, and open the replays behind them.
Named track, page and screen events now sit beside errors, traces and replay. Opt in per project (off by default), then use Talaria.analytics after a CMP. Included on the plan — not a separate SKU.
@newtalaria/browser (errors, replay, vitals), @newtalaria/react (ErrorBoundary and React Router), @newtalaria/nextjs (client, server and edge entrypoints) and @newtalaria/node (process + HTTP traces). Tracing stays off until you enable it.
Three Composer packages: talaria/talaria (core), talaria/silverstripe and talaria/laravel. Laravel covers exceptions, Eloquent, queues, HTTP client, Artisan, Octane, Horizon and Livewire.
talaria 0.2.3, talaria_flutter 0.1.3 and talaria_serverpod 0.1.3 are on pub.dev. Docs cover install, configuration, capture, tracing, HTTP, Flutter navigation and Serverpod Relic, Postgres and FutureCalls — plus a checkout tutorial.
Official SDKs capture transactions and child spans. Turn tracing on under Project settings. Error transactions are always sampled; successful requests follow the project rate (10% until you change it).
Transaction waterfalls, RED metrics, DB and HTTP spans, Web Vitals and release health in the dashboard — straight from the official SDKs, with no agent to install.
Events carry breadcrumbs and user-impact context so you can see who hit an issue and what happened before it. Paid plans notify via webhook and Slack alongside email.
Public launch — event ingest, server-side fingerprinting, issue triage, alerts, organisations, dual-meter pricing and the marketing site.
Open, resolved and ignored statuses with event counts, first and last seen, and severity in the dashboard.
Events are grouped into issues via server-side fingerprints — clients never compute identity.
Create alert rules and receive notifications when issues match.
Multi-project workspaces, team roles and project-scoped ingest keys.
Developer is $5 per month. Connect an agent, or open the dashboard and send the first event.