Skip to content
bucker

Comparison

What we do that the category does not

“Has AI” is table stakes now — every incumbent shipped it. The question worth asking is narrower: when the AI says it fixed something, what exactly are you being asked to believe, and can you check it without trusting the vendor who is billing you for the answer?

How to read this page

The differences

How Bucker differs from the category norm
DimensionThe category normBucker
What "verified" meansA model asserts the fix is correct, and a pull request appears. The claim is a sentence in a description field.A repro test that failed before the patch and passed after it, a suite result, and a verification tier that an audit function refuses to inflate. Tier A requires a full suite that passed; tier C is what you get when there is no passing suite.
Whether you can check it yourselfVerification, where it exists, is something the vendor did and reports on. There is nothing to re-run.The evidence bundle is signed Ed25519 over DSSE and carries an RFC 6962 Merkle inclusion proof. The bucker-verify CLI re-derives every claim offline, makes no network calls, and can check the signature against a key you supply rather than the one inside the document.
Where you approveYou leave your editor, open a dashboard, find the item, and click. MCP servers, where they exist, are read surfaces.The MCP server implements input_required, the standards-track human-approval primitive, across its write tools — and renders the approval and evidence cards as MCP Apps when the client negotiates for them. You answer the prompt in Claude Code.
Payload shape for agentsHuman-shaped JSON served through an MCP wrapper: whole stack traces, unbounded breadcrumb arrays, no published size.A bounded projection under 500 estimated tokens with the count published on every response, a reduction ladder that reports what it dropped, hydration handles to fetch the rest deliberately, and a pinnable schema version.
AI application failuresTraces and evaluation scores in a separate observability product, with their own vocabulary and their own inbox.Detectors for tool-schema violations, retry loops and silent truncation call the same ingest function your crashes do. One fingerprint, one lifecycle, one alerting path, one remediation loop.
Agent permissionsAn organization-level AI toggle, and a setting for whether the bot may open a PR.Agents are principals with sponsors, RFC 8693 delegation chains, budgets scoped to (agent, human, incident) that reserve before each generation, incident-bound blast-radius tokens, action-granular audit, and automatic token revocation on anomalous behaviour.
Who can mergeConfigurable. Auto-merge is often a roadmap item or an opt-in setting.A human, unconditionally. Refused to non-human principals before policy is consulted, the scope filtered out of every agent token, the trust ladder pinned so no rung reaches it — and no merge method in the source-control client to call.
Single sign-onFrequently gated behind the enterprise tier; the "SSO tax" is a documented churn trigger.On every paid tier, alongside SCIM — including SCIM for agent principals. A test fails the build if a paid tier stops including it.
SeatsPer-seat pricing, sometimes per contributor, so the bill grows as the team does.Unlimited human seats on every tier. There is no seat-count field in the plan schema; the type is the literal string "unlimited". Agent principals are unlimited on every tier too: what costs money is the verified-fix lane, and that is metered per outcome.
A spike in trafficMetered per event, so an incident is also an invoice. Spike protection, where it exists, may stop ingesting.Over-quota events cost nothing and degrade to deterministic stratified sampling, with every degraded event receipted. A storm from one deploy is one issue and can produce at most one billable fix.
Unused creditsCommonly expire at the end of the billing period.Roll over on every tier, spent before you are metered — capped at three months of your allotment, which we state rather than bury.
The priceFrequently "contact us" from the first paid tier upward.Published, and read at build time from the same public endpoint the product reads when it enforces your quota, so the page cannot drift from the ladder.

If you already run Sentry

Bucker’s ingest speaks Sentry’s envelope grammar, so an unmodified Sentry SDK pointed at a Bucker DSN reports errors here. The change is one string, and nothing is installed. That makes it the cheapest way to evaluate Bucker against your own errors, and it is an evaluation path rather than a full replacement: the table below is exactly which of the things a Sentry SDK sends are consumed today.

An envelope is never refused for carrying an item Bucker does not consume. The error in it is stored, and each unconsumed item is dropped with a named outcome you can see on the drops panel, so nothing is lost silently.

Which Sentry item types Bucker consumes today
Item typeConsumedWhat happens today
eventYesErrors and messages: stored, grouped, symbolicated where Debug IDs allow, turned into an essence
sessionYesPer-session release health: crash-free sessions and users
client_reportYesAccounted at the accept edge, so what your SDK discarded shows beside what Bucker dropped
sessionsPartlyAggregate session buckets (request-mode sessions; the Node, Python and other server SDKs) feed crash-free sessions and adoption. Crash-free users is shown as "not derivable from aggregates" rather than guessed: an aggregate carries no per-session identity
transactionYesStored as spans in the same table OTLP spans land in, named after the transaction; metered per span, as OTLP is
spanYesBucker's own span shape, and Sentry JS v11's streamed span batches
logYesBucker's own log shape, and Sentry's log record (level, typed attributes, environment, release)
check_inYesRecorded on the cron monitor named by monitor_slug; a monitor is created from monitor_config when it does not exist yet
feedbackYesStored as feedback (also user_report), linked to the issue of the event it names, with the widget route's spam scoring
attachmentPartlyText and logs (scrubbed first) and minidumps are stored and shown on the event's Attachments panel; screenshots only where the project opts in. Other binaries and credential-named files are refused by name, not billed
profileRefusedRefused by name (unsupported_item_type:profile, also profile_chunk), not billed. Bucker has no profiler; set the profile sample rate to zero
replay_recordingPartlyBucker consumes its own recorder's format. A Sentry Replay (rrweb replay_event + replay_recording) is refused by name as unsupported_item_type:sentry_rrweb_replay, not billed
state_patchesNot yetNot a Sentry item: a Bucker SDK item the worker does not consume yet
product_eventYesNot a Sentry item: a Bucker SDK item for product analytics. Accepted for a project that has switched product analytics on, metered as its own category (0.05 units an event, whoever the user is), answered per item, and folded into sessions and rollups; no raw event is kept. Where the deployment has not enabled the product_events pillar the worker defers it with the reason behind pillar

Where a competitor is genuinely ahead