What's new in Squasher.
New features, improvements, and fixes, written up from the changes we ship to production.
Sign up with Google or GitHub, and see what each release shipped
Open sign-up with more ways in, release build details on service pages, and your own model key for Squasher Agent.
- NewSign up with Google or GitHub. Email sign-up is open again.
- NewService and release pages show the version, commit, branch, and platform of each release, read from your telemetry. Docs
- NewUse your own OpenRouter key for Squasher Agent. Its calls then use no AI credits. Docs
- NewAgent chat history shows where each thread started.
- ImprovedUndo an issue merge from the notice that confirms it.
- ImprovedIssue titles are clearer, and generic issues show their HTTP request details.
- ImprovedThe Logs Volume view keeps its controls and totals while you change the grouping.
- FixedMonitors behind an HTTPS proxy no longer open false SSL certificate incidents.
- FixedAn SDK error sent twice no longer creates two issue events.
- FixedOAuth clients receive only the scopes they ask for.
Sentry SDKs work as is, and AI calls count once
Point an existing Sentry SDK or sentry-cli at Squasher, read more AI span formats, and get a redesigned incident page.
- NewSquasher accepts every Sentry SDK item type and the numeric DSN, so existing Sentry SDKs send without code changes. Docs
- Newsentry-cli uploads source maps, creates releases and deploys, and sends cron check-ins to Squasher. Docs
- NewSentry cron check-ins signal the heartbeat monitor with the same slug.
- NewAI observability reads more OpenTelemetry AI span kinds, AI log events, and Claude API spans, and counts each model call once. Docs
- NewUsage forecasts and guardrails on the billing page, and monthly AI credits for each organization. Docs
- NewUptime checks from several locations, with N-of-M confirmation and failure evidence. Docs
- NewBounded regex search for logs and events.
- NewA redesigned incident detail page.
- ImprovedThe Pro plan includes twice the telemetry, with no query fees.
- ImprovedThe notification channel picker is easier to scan.
- ImprovedSquasher Agent chat streams answers and approval buttons more reliably.
- FixedWeb Vitals from the newest Sentry browser SDK show on the Vitals page.
- FixedInvestigations no longer look stuck in the investigating state.
Services, anomalies, reliability reports, and Terraform
A service map with releases, anomaly detection, scheduled reliability reports, a Terraform provider, and Microsoft Teams alerts.
- NewServices, a service map, and releases, found in the traces and logs you already send. Docs
- NewAnomalies pages, with a shared detection engine, backtests, and an alert builder. Docs
- NewReliability reports for incidents, monitors, on-call, and issues, with MTTD, SSL expiry, and a weekly summary email. Docs
- NewA Terraform provider for monitors, status pages, notification routing, dashboards, and on-call. Docs
- NewMicrosoft Teams as a notification destination, with Adaptive Cards. Docs
- NewRequest latency as a monitor source.
- NewSource maps apply to stack traces, and grouping rules control how errors become issues. Docs
- NewInvestigations with a live timeline, and actions that run only after you approve them. Docs
- ImprovedChange an incident's status from the incidents list.
- FixedOrganization retention overrides apply at ingest.
Faster log search, an account inbox, and readable AI conversations
Log and trace rails answer faster on 30-day windows, notifications collect in one inbox, and AI request pages read like a conversation.
- NewAn account inbox that collects your personal notifications in one place.
- NewAI request and trace pages show the conversation, tool calls, usage, and provider errors, with highlighted code. Docs
- ImprovedFast common-word log search and faster log and trace rails on 30-day windows.
- ImprovedThe Logs column picker offers the attributes your rows carry.
- ImprovedSquasher reads current and older OpenTelemetry attribute names, so older instrumentation still matches.
- FixedSession replay uploads no longer drop chunks when they arrive together.
- FixedSecrets in relative URLs are redacted from session replays.
One Squasher Agent, a redesigned dashboard, and status pages on squasher.ai
Every agent is now one Squasher Agent, every dashboard page uses one layout, and status pages can live at squasher.ai/status.
- NewOne Squasher Agent in the dashboard and in Slack's assistant pane. It shows its steps and asks before it writes. Docs
- NewStatus pages at squasher.ai/status/<slug>, with Squasher branding and subscriptions. Docs
- NewUptime checks can run every 30 seconds.
- NewLink more than one GitHub repository to a project.
- NewSession replay is rebuilt around one session story that links the related traces.
- ImprovedEvery dashboard page is redesigned on one layout, and the dashboard opens your last project.
- ImprovedIssue bulk actions, incident reopen, short issue IDs, and swipe navigation on phones.
- FixedLogs from Heroku, Fly.io, CloudWatch, and Cloudflare Logpush are stored correctly.
- NewThreaded incident comments with @mentions and notifications.
- ImprovedComment timestamps show the exact time on hover.
- FixedThe project and organization switchers in the header go to the right page.
AI observability, rebuilt
New AI observability pages for traces, sessions, and single model calls, with simpler Vercel AI SDK setup.
Post-incident review links on status pages
Publish a link to your post-incident review on the status page.
- NewStatus page incidents can link a post-incident review. Docs
- ImprovedAI triage holds back an automatic fix when it has no evidence for it.
- ImprovedSimpler copy actions in the log viewer.
Go squash some bugs.
Connect one service and send one real event. The crew takes it from there.
Free plan. No credit card required.