Read-only first · start free · no card

Slack already contains the work. Keep the thread without maintaining another dashboard.

Select the channels that matter and run a read-only Shadow Report. See the commitments, owner gaps, and open questions it would have caught. Nothing is posted or changed. Save and schedule the brief only when the result earns your trust.

0 writes in Shadow Mode 5 workflow contracts 08:00 scheduled delivery P-256 signed webhooks
Workerv6.7.0 / live
Tool surface20 Slack read/write + 6 AI workflow
Contracts5 validated shapes
DeliveryP-256 signed
Analyticscontent-free
OSS receipt20,703 npm downloads from Jan 1 through Aug 17, 2026

The problem is not search. It is rebuilding the state of work.

The decisions, risks, owners, and open questions are already in Slack. The expensive part is rebuilding operating state from them every morning.

Manual reconstruction Search → skim → copy → reshape → chase owners

A person rebuilds the same brief from raw messages, then fixes the output shape before another system can use it.

Continuity loop Observe → review → save → schedule

Start with a read-only report. When the findings are useful, keep the channels, contract, cadence, and destination so the next brief starts from the same operating definition.

Live service snapshot.

This reads the deployed worker's status contract on load — so the proof stays tied to the running service, not to copy someone forgot to update.

Checking live worker status…
Open /status
Hosted version Current deploy, same-origin
Free Slack tools plus hosted inference tools
Token modes Ephemeral by default, persistent opt-in
Account path /account Usage, billing, reconnect, config

Reads /status directly from the worker and keeps deployment review plus support on the hosted surface.

Start with evidence. Add automation in measured steps.

The first run is read-only. Save the operating contract only after the findings are useful, then add a schedule or signed delivery when the workflow has earned that responsibility.

Run a Shadow Report.

continuity_shadow_report observes one to five channels and returns a validated report. It cannot post or deliver anything.

Save the workflow that proved useful.

Keep the selected channels, output contract, timezone, cadence, and destination under one named profile.

Deliver it at 8am.

Run daily or every Monday in the workflow's IANA timezone. Idempotency keeps one schedule slot to one Slack delivery.

Send a signed incident/status payload downstream.

Pro and Team can deliver P-256-signed webhooks so the receiving system can verify the payload before accepting it.

Hosted earns its keep when the workflow has to run tomorrow.

Self-host when runtime ownership matters most. Choose hosted when running inference, retaining the workflow definition, scheduling the brief, and verifying downstream delivery should be the product's job.

Start Free against one real workspace. Move to Pro only when the workflow becomes part of daily operations.

SELF-HOST runtime ownership yours token rotation yours scheduling yours HOSTED PRO credential mode persistent / opt-in AI workflow briefs included 08:00 delivery daily or Monday

Four operating workflows, mapped to profiles.

support_inbox for support triage, incident_room for live incidents, exec_brief for founder leverage, product_launch_watch for product ops.

Free to try. $19/mo when it's a daily habit.

Free has real monthly quotas — 2,000 requests + 25 AI tool calls. Individual adds unlimited use, saved schedules, signed delivery, and Slack OAuth with a read-only default. Small team is flat for up to 5 operators.

Free
$0

All 26 tools — 20 Slack read/write + 6 AI (25 calls / mo) + 2,000 requests / mo. 1 workspace. No card.

Start free
Small team
$49/mo

Individual + shared workflow profiles + up to 5 operators + 24h support. Flat, not per-seat. $490/yr annual.

Start small team
Safeguard · waitlist
$199/mo (waitlist)

Agent approval gates, searchable workspace memory, and workspace-wide audit — still in development. The live scheduling layer stays in Pro and Team.

Join the waitlist

Annual billing saves two months on Pro and Team. Ops engagement (SLA, SOC2 evidence, multi-tenant isolation) is quoted per organization — talk to us.

Evidence, not theatre.

Token handling, delivery verification, analytics, and live status stay inspectable. No Slack content enters product analytics.

Token & storage

  • Ephemeral by default; persistent only with explicit consent
  • AES-256-GCM encryption at rest
  • Slack responses stay on the live request path — no message telemetry stored

Analytics

  • Cloudflare Web Analytics on hosted web pages only
  • Separate from Slack MCP request handling
  • Documented on the privacy page

Buyer review

  • Pricing, security, and privacy are all hosted-native
  • Live status stays visible at /status
  • Security is the controls surface; deployment review is the higher-touch path

Automate the brief, then keep the boundary visible.

Scheduled delivery is live live

  • A saved workflow profile can run daily at 8am or Monday at 8am in its IANA timezone
  • Slack delivery is idempotent, so one schedule slot posts once
  • Pro and Team can also deliver a P-256-signed incident/status webhook

Workspace memory in development

  • Searchable memory across the workspace's history
  • Part of the Safeguard tier — on the waitlist
  • Nothing here is billable until it ships

Agent approval gates in development

  • A human approves before an agent acts on the workspace
  • The reason Safeguard exists — join the waitlist
  • Price-locked at $199/mo while it's built

Buyer objections, answered directly.

What does the free tier actually include?

All 26 tools: 20 Slack read/write, plus the 6 AI tools metered at 25 calls per month — one shared pool across slack_catch_me_up, slack_triage, slack_smart_search, slack_workflow_brief, continuity_shadow_report, continuity_asana_discrepancy_report that resets on the 1st. 2,000 requests per month. One workspace. Published quotas. No card.

What does Pro at $19/mo add over free?

Unlimited requests, unlimited AI tool calls, saved profiles, signed webhooks, Slack OAuth with a read-only default, optional scheduled delivery, and 1 Slack workspace.

Why $49 flat for Team?

Small teams budget as line items, not per-seat sprawl. The flat plan covers up to 5 operators in one shared Slack workspace.

When should I stay self-hosted instead?

Stay self-hosted if your team wants full runtime ownership. The MIT-licensed package ships all 21 tools — 20,703 npm downloads from Jan 1 through Aug 17, 2026. Hosted adds Workers AI inference, saved schedules, signed delivery, and opt-in encrypted credential storage. Slack may still rotate session credentials on either path.

What about analytics and retention?

Hosted web pages use Cloudflare Web Analytics for aggregate metrics. Slack responses stay on the live request path and are not retained as message telemetry.

See what would have been caught.

Pick the channels, choose the output contract, and run the first report read-only. No card, no post back to Slack, and no new dashboard to maintain.