Ch 01 · Features
BulkReach is not a bigger textbox. It is the queue, the pacing, the consent checks and the audit trail that turn one message into hundreds of delivered ones — without you watching it happen.
A standard OAuth install, a synced member directory, and sending that respects the rate limit Slack actually enforces.
The regular “Add to Slack” flow, requesting four bot scopes and nothing more: team:read, users:read, users:read.email and chat:write. Install state is signed and tied to your browser session, so an install link cannot be replayed into someone else’s workspace.
Pull the full workspace directory on demand, paginated automatically. Bots, deactivated accounts and Slackbot are filtered out. People who leave are marked inactive rather than deleted, so historical delivery logs keep pointing at a real person.
Slack allows roughly one chat.postMessage per second per workspace. BulkReach paces to that automatically and backs off when Slack returns a 429, so a large send completes instead of half-failing.
Messages go to people, not to a channel everyone has to mute. DMs are opened implicitly at send time, so there is no setup per recipient.
QR pairing, a session that survives restarts, and consent enforcement that is not optional. WhatsApp automation is unofficial — we build accordingly.
The same flow as linking WhatsApp Web: open Linked devices on your phone and scan. The pairing screen updates live as the session connects.
Session credentials are encrypted with AES-256-GCM and stored, so a deployment or restart resumes the connection instead of demanding a fresh scan every morning.
Messages are spaced three to eight seconds apart at random, on top of a hard floor enforced inside the sending service itself. A predictable cadence looks automated, and looking automated is what gets numbers banned.
Once when you pick the audience, and again at the moment of delivery. Someone who opts out between queueing and sending never receives the message.
Slack members, WhatsApp contacts and imported numbers live in one place, searchable and segmentable together.
Everyone you can reach in a single table, tagged by where they came from — Slack, WhatsApp or CSV — so you never have to remember which list someone is on.
Common header names are recognised automatically (phone, number, mobile, whatsapp). Numbers are normalised to international format, duplicates merged, unreadable rows reported. Nothing is marked opted-in unless you explicitly confirm the list agreed to be contacted.
Tag people once and reuse the tag as a segment when composing. Bulk-tag a selection in one action rather than editing rows one at a time.
Search across names, emails and phone numbers; filter by source or tag. Selection survives filtering, so you can build an audience across several searches.
Record or withdraw consent, add tags, or delete — for hundreds of recipients at once, from one selection bar.
Every WhatsApp and CSV recipient shows their opt-in state in the list, so you know what a send will actually reach before you write a word.
Templating, live preview and channel selection, with the mistakes caught before they ship rather than after.
Use {{name}}, {{first_name}}, {{email}} and {{phone}}. Each is resolved for the individual recipient at send time and stored with the delivery, so the log shows exactly what each person received.
See the message rendered as a real selected recipient would receive it, not as a template with braces in it.
A placeholder that does not match a known variable is flagged before you send, and left visible in the message rather than silently replaced with a blank.
Select Slack, WhatsApp or both. Each recipient is routed on the channel that can actually reach them, and the summary tells you how many will be skipped and why.
Scheduling is not a separate mode. It is the same queue, held until it is due.
Pick a date and time when composing. A scheduler checks every minute and dispatches anything that has come due.
Change the message or the time right up to the moment the send begins. Editing the text re-renders every queued recipient, so nobody receives the previous draft.
Each workspace sets its own timezone, and scheduled times are displayed in it — so a team spread across three countries reads the same schedule the same way.
A send is not a fire-and-forget. Every recipient is a row you can look at afterwards.
Sent, failed, pending or skipped — for every recipient on every channel, with the exact text delivered and the provider error where one occurred.
Each recipient is queued separately, so an interrupted send resumes where it stopped. Nobody gets messaged twice because a browser tab closed.
How many messages went out this month, split by channel, on the overview — the foundation for plan limits and for answering “how much are we sending?”.
Isolation enforced by the database, secrets encrypted at rest, and an audit you can run yourself.
Every table is scoped by workspace and guarded by Postgres RLS. A query cannot cross workspaces even if the application layer has a bug, because the database refuses to return the rows.
Slack bot tokens and WhatsApp session credentials are AES-256-GCM encrypted before storage. The Slack token column is additionally revoked from client roles, since row-level security does not protect individual columns.
Only three server-side paths bypass row security — the send dispatcher, the OAuth callback and the WhatsApp service — and each scopes every query by workspace explicitly.
A single accent on a neutral base, tuned separately for both themes, with reduced-motion respected throughout. It defaults to light and remembers what you pick.