Highlight.io

Session replay and error tracking for full-stack debugging

3.2(0 reviews)freemiumFounded 2020United States

free tieropen sourceself hostableapi access

Integrates with

Open-source full-stack monitoring (session replay, error monitoring, logging, tracing), self-hostable via Docker. Acquired by LaunchDarkly in April 2025; cloud accounts are being migrated to LaunchDarkly Observability.

What Highlight.io actually is

Highlight.io is an open-source, full-stack monitoring platform that combines session replay, error monitoring, log management, and distributed tracing in one product, and it is currently mid-flight into someone else's platform. LaunchDarkly announced the acquisition of Highlight (Highlight Inc., Seattle) on 23 April 2025; the co-founders, Jay Khatri and Vadim Korolik, moved over to lead observability there, and by November 2025 the company published a migration guide titled "Migrating from Highlight.io to LaunchDarkly Observability." Visit highlight.io today and you are served LaunchDarkly's marketing site (the footer reads Catamorphic Co., LaunchDarkly's legal entity); the surviving product surfaces carry a banner telling account holders to migrate. Any review that describes this as a normally available standalone vendor in late 2026 is describing last year's product.

That said, the thing itself is real and worth understanding. On the product level, Highlight records user sessions with console and network capture and privacy controls, groups application errors with stack traces and customizable alerting rules, ingests logs with searchable dashboards and performance visualizations, and follows requests across services with traces — the instrumentation is powered by OpenTelemetry, which is the right architectural choice and makes the SDK story broad (the old landing page listed support across the usual modern frameworks plus integrations for issue trackers and analytics tools). The design bet was correlation: one click from an error to the exact user session that hit it, with the logs and traces for that request attached. That is genuinely the debugging workflow teams want, and Highlight was one of the few open-source options delivering it as a single stack.

The self-hosting story was always the differentiator: the entire platform runs from the public repository (github.com/highlight/highlight) via Docker — the documented hobby path is a clone with submodules followed by ./run-hobby.sh — with Dev, Hobby, and Enterprise self-host variants documented. Session data, errors, logs, and traces stay on your infrastructure, which is the argument that survives the acquisition intact: the source is still open and the community is invited to maintain it. LaunchDarkly's own FAQ at the time of the deal said Highlight "will remain open source in its current state," with critical fixes released back to the repository and a final support date to be announced. As of our review, that final date is the open question.

Pricing: what you really pay

The published cloud pricing (as last served on highlight.io/pricing and mirrored by third-party plan trackers) has four tiers. Free at $0/month: 500 monthly sessions, AI error grouping, up to 15 seats, with quota figures for errors, logs, and traces that trackers list as roughly 1,000 errors, 1 million log records, and 25 million trace spans per month. Pay-as-you-go at $50/month base: a small bundle of included usage across sessions, errors, logs, and traces with per-unit overages, 3 dashboards, 2 projects, up to 15 seats, and short retention (7 days on the trackers we read; one plan aggregator instead lists 3 months for sessions and errors — the figures disagree, so verify before budgeting). Business at $800/month: unlimited dashboards, projects, and seats, data-ingest filters, and custom retention policies. Enterprise is custom-quoted with SAML/SSO, RBAC, audit logs, custom MSAs and SLAs, and data export.

Two things stand out about that ladder. First, the free tier was genuinely usable — 500 sessions and up to 15 seats covers a small product team completely, which is more generous on seats than most rivals. Second, the step from free to $50/month base plus overages to $800/month is steep and the middle of the market is thin; third-party reviewers comparing Highlight with Sentry for solo developers reached the same conclusion about the jump. Self-hosting is the pressure valve: run the hobby stack yourself and the software license costs nothing, at the price of operating ClickHouse, Postgres, Kafka, and the rest of the stack yourself.

Our caution on all of these numbers: they describe a product in the process of being folded into LaunchDarkly's commercial platform, where pricing is LaunchDarkly's pricing. Treat published Highlight tiers as the last known state, not a quote you can accept today.

Performance & infrastructure

Architecturally Highlight is the modern observability shape: OpenTelemetry-compatible ingest, a GraphQL backend, ClickHouse-class storage for high-volume events, and a frontend that searches across millions of sessions. Session replay captures console output, network requests, and DOM state with configurable privacy masking; error grouping combines AI grouping with customizable rules; logging supports dashboards and performance visualizations. Users reported fast search across large session volumes, and the docs are unusually candid for a vendor — they include architecture write-ups, a competitors page, changelogs, and telemetry disclosure.

For the hosted product, Highlight published no data-center map, region picker, or infrastructure SLA that we could find, and the acquisition makes that harder to pin down rather than easier: the infrastructure runs today as it did, with LaunchDarkly stating at deal time that Highlight's existing infrastructure and site would run alongside LaunchDarkly until full integration. Whether data residency options, regional hosting, or compliance certifications carry over to LaunchDarkly Observability is a question for LaunchDarkly's trust documentation, not Highlight's.

Self-hosted, the infrastructure is whatever you provide. The hobby stack is Docker-based and reasonable for a small team; the enterprise self-host variant exists for larger installs. This is where the product's data-residency story is strongest: your sessions never leave your own servers, and GDPR questions answer themselves. The cost is operational: this is a multi-service stack (database, queue, storage, collectors), and upgrades, backups, and capacity are your problem.

Where it falls short

The first and largest issue is product continuity. Independent Highlight cloud accounts are being migrated to LaunchDarkly; the domain serves LaunchDarkly's site; the open-source repository's support horizon is defined as "until we tell you the final date." Picking Highlight today as a new standalone monitoring vendor means betting on a migration path and a new commercial relationship you have not chosen. Existing users should evaluate the LaunchDarkly Observability terms before their next renewal.

Second, the pricing structure has real gaps: nothing between a small free tier and $800/month, conflicting published quota and retention figures across trackers, and usage overages that apply per data type (sessions, errors, logs, traces) — the four dimensions where bill shock happens. The product also always leaned web-first: mobile session replay and native SDK depth lagged dedicated players.

Third, self-hosting is freedom with homework. The hobby stack is fine for evaluation, but running the production-grade version means operating several stateful systems; there is no managed "self-hosted" middle tier. And with the founding team now inside LaunchDarkly, community maintenance of the repository is exactly what LaunchDarkly's stated plan implies — best-effort, with critical fixes flowing from one company's roadmap.

Finally, on evidence standards: Highlight published no public uptime history, no third-party audit results we could verify, and no benchmark figures. Our usual advice for monitoring tools applies with extra force here — instrument a pilot before you standardize.

Who should pick this

We would not pick Highlight.io today as a new cloud customer; you would effectively be signing up for LaunchDarkly Observability, and you should evaluate that product directly on its own terms, against Sentry, LogRocket, and PostHog. If you are already a Highlight customer, the decision is simpler: complete the migration deliberately, check what your quota and retention become, and confirm the data-processing terms before the deadline your account is given.

Where Highlight still earns a look is self-hosting. If you want open-source session replay tied to error tracking, logs, and traces in one stack, on your own metal, with no vendor touching user data, the repository remains the most complete option of its kind — we know of no other open-source project covering all four signals together at comparable polish. Teams with compliance constraints (healthcare, fintech, public sector) that ruled out third-party session capture are exactly the audience. Budget for the operational cost, or wait and see what LaunchDarkly does with the open-source core.

Key takeaways

Highlight was acquired by LaunchDarkly on 23 April 2025; highlight.io now serves LaunchDarkly's site and cloud accounts are being migrated to LaunchDarkly Observability.
The product combined session replay, error monitoring, logging, and tracing with OpenTelemetry-based instrumentation — one of the few open-source, self-hostable stacks covering all four.
Last published cloud pricing: Free $0/month (500 sessions, up to 15 seats), Pay-as-you-go $50/month base with per-unit overages, Business $800/month, Enterprise custom; third-party trackers disagree on quota and retention details.
Self-hosting runs from github.com/highlight/highlight via Docker (Dev/Hobby/Enterprise variants); the repo stays open source, but a final community support date has been announced as forthcoming and not delivered.
No published infrastructure SLA, region map, or uptime record for the hosted product; self-hosted is where the data-residency case is strongest.
Our view: pick the self-hosted stack if you need open-source full-stack monitoring on your own infrastructure; for hosted monitoring, evaluate LaunchDarkly Observability or rivals directly instead of the Highlight brand.

Sources

Pros

One stack for session replay, errors, logs, and traces with OpenTelemetry-based instrumentation and cross-signal correlation
Genuinely usable free tier: 500 monthly sessions and up to 15 seats at $0/month
Fully self-hostable from the public repository (Docker hobby stack), keeping session data on your own infrastructure
Strong docs for a vendor: architecture notes, self-hosting guides, changelogs, telemetry disclosure
Acquired into LaunchDarkly with the founders leading observability there, so the hosted product has a defined continuation path

Cons

Standalone product is being absorbed: highlight.io serves LaunchDarkly's site and cloud accounts are being migrated
Open-source repository's final community support date is still unannounced, so self-hosters face an uncertain maintenance horizon
Thin pricing ladder: nothing between the free tier and $800/month Business, with per-type usage overages and conflicting published quotas
No published infrastructure SLA, region map, or uptime record for the hosted product
Self-hosting means operating a multi-service stack (database, queue, collectors) with no managed middle tier

Best For

Teams that want open-source, self-hosted session replay plus error, log, and trace monitoring on their own infrastructure — and existing Highlight cloud customers deciding whether to complete the LaunchDarkly migration.