PRODUCT // OPENRIFT // v1

What
Openrift does.

AI coworkers that live in your Slack and WhatsApp, run real code in their own compute environment, connect to 3,000+ tools, and ship the work — not a chat reply.

3,000+
Integrations
2 wks
Discover → Deploy
Slack / WA
Where agents live
SOC 2
Type 1 today
01 / OUTPUTS

Agents that ship outputs, not text.

PDFs · dashboards · web apps · code

Most AI tools explain the work.
Openrift does the work.

Ad-spend audit
Chat-only AITells you how to audit your ad spend.
Openrift →Audits it. Pulls Meta + Google. Hands you the PDF.
Internal dashboard
Chat-only AIWrites pseudocode for a dashboard.
Openrift →Builds it. Hosts it. Sends you the URL.
Customer report
Chat-only AIOutlines the structure of the report.
Openrift →Queries Stripe + HubSpot. Renders the PDF. Emails it.
Bug triage
Chat-only AISuggests how to triage.
Openrift →Reads support, dedupes, files Linear tickets with repro.
// example output · monday pulse
Openrift · Daily Pulse
14 MAY · 09:00
MRR
184,500SGD
+18%
Pipeline
1.2MSGD
+62%
Burn
92kSGD
flat
Runway
14months
–1mo
// narrative

May closed up 18% MoM — driven by 4 mid-market closes in SG/MY. Burn held flat despite the ops hire. Pipeline weighted at SGD 1.2M with three enterprise in legal review. Runway shortens by a month from April; conservative new-deals model puts it back at 16 by July.

// actions queued
  • Confirm Antler SEA intro for warm pipeline (4 names queued)
  • Forecast refresh scheduled Wed — pending close updates from CRM
  • Audit trigger: Meta-TOFU-V4 dropped to 0.7% CTR yesterday
Sources — Stripe · HubSpot · Sheets · Headerv 12 · auto
02 / PRESENCE

Lives in Slack and WhatsApp.

Persistent · mentionable · always there

Agents are not a tab you open. They live where your team already talks. You mention them, they work. They follow threads, ask clarifying questions, hand back finished outputs in the same channel.

// sample thread · leadership channel
#leadership
Openrift · online
Mei Lin
Mei Lin9:02 AM
@Openrift audit our Meta Ads spend last month vs. CAC by campaign — flag anything that's burning.
O
Openrift App9:04 AM
Pulled Meta + Stripe for May. Total spend SGD 38,420 across 7 campaigns. Three are torching budget — full breakdown attached.
// AD-SPEND AUDIT · MAY
3 campaigns flagged · est. SGD 11,200/mo wasted
  • TOFU-IG-Reels-3 — CPL 4.2× ICP avg, ROAS 0.62
  • Retarget-V2 — frequency 9.4, declining CTR week 3+
  • Lookalike-1% — CAC SGD 312 vs SGD 95 target
PDFMeta-Ads-May-2026-Audit.pdf
Mei Lin
Mei Lin9:05 AM
Kill the first two. Pause Lookalike-1% and rebuild it.
O
Openrift App9:05 AM
Paused both in Meta Ads Manager. Logged the change in #ops-log. New Lookalike audience brief drafted — review in DM.
Message #leadership
Workspace-native

Slack

Mention the agent in any channel. It picks up context from threads, reads files, replies in line.

  • Public + private channels
  • Thread-aware replies
  • File + image attachments
  • Slash commands and shortcuts
Where SEA actually works

WhatsApp

Half your team runs on WhatsApp. Agents reply in business groups, accept media, send PDFs back.

  • Business API integration
  • Group + DM threads
  • Voice notes (transcribed)
  • Media + document replies
Enterprise overlay

Microsoft Teams

For regulated or enterprise environments. Same agent, same context, surfaces in Teams channels.

  • Channel mentions
  • Adaptive cards for approvals
  • Tenant-scoped install
  • Compliance-mode logging
// SEA · WhatsApp-native

Half your SEA team
lives on WhatsApp.

Founders on the road. Ops leads in the warehouse. Vendors who never installed Slack. Agents reach them where they already are — business groups, DMs, voice notes, document replies.

O
Openrift · Bookkeeping
via Openrift · last active just now
// AP · DAILY
3 new invoices arrived this morning. Read and queued. Two match POs cleanly — flagged one anomaly.
10:02
What's the flag?
10:03
Cloud Storage SGD 1,840 — 32% higher than monthly average. Worth a look before approval.
10:03
Type a message…
03 / COMPUTE

Each agent has its own computer.

Isolated · sandboxed · real

Not a chatbot
with a typing cursor.

Each Openrift agent runs in a dedicated cloud sandbox. It executes real code. Hits your tools through real APIs. Renders real outputs. The Slack reply is the surface — the work happens underneath.

Workspace boundary is a hard technical boundary. Your compute, your data, your secrets. Never mixed.

// 01
Isolated cloud sandbox

Every engagement gets a dedicated compute environment. No data crosses tenants.

// 02
Executes real code

Python, Node, shell. Installs packages. Runs SQL. Hits APIs. Builds and ships artefacts.

// 03
Headless browser

Logs into web UIs when an API doesn't exist. Scrapes, fills forms, downloads files.

// 04
Filesystem and storage

Writes PDFs, spreadsheets, charts. Persists working state across sessions.

// 05
Hosting and deploys

Spins up internal tools and dashboards. Returns a live URL your team can visit.

// 06
Long-running jobs

Audits, exports, report builds that run minutes to hours. Pings back when ready.

04 / TOOLS

Connects to 3,000+ tools.

Native + custom connectors

If your stack has it, agents probably can too. Native integrations for the big platforms; custom connectors built during deploy for the regional, niche, or internal systems.

Sales & CRM
  • Salesforce
  • HubSpot
  • Apollo
  • Pipedrive
  • Attio
Marketing & Ads
  • Meta Ads
  • Google Ads
  • Instantly
  • Customer.io
  • Klaviyo
Engineering
  • GitHub
  • Linear
  • Jira
  • Sentry
  • PostHog
Finance & Ops
  • Stripe
  • Xero
  • QuickBooks
  • Brex
  • Mercury
Productivity
  • Notion
  • Google Drive
  • Airtable
  • Coda
  • Dropbox
Communication
  • Slack
  • WhatsApp
  • Microsoft Teams
  • Gmail
  • Outlook
E-commerce
  • Shopify
  • BigCommerce
  • WooCommerce
  • Lazada
  • Shopee
And 2,950+ more →Browse the full grid
05 / MEMORY

Learns your business.

Gets sharper every month
// 01

Channel context

Reads the channels you invite it into. Picks up vocabulary, names, recurring questions.

// 02

Internal skills

Each repeatable workflow becomes a documented skill the agent can re-run on command.

// 03

Living knowledge base

Decisions, naming, conventions — captured. Searchable. Auditable.

// 04

Workflow playbooks

Multi-step procedures with confirmation gates. Versioned. You can read what it will do.

06 / CONTROL

Stays in your control.

Confirmations · audit · pause

Agents are eager. We make sure they ask. Anything irreversible — external sends, prod deploys, money moves — pauses for human confirmation. You always know what happened, and you can stop it instantly.

Confirmation gates

Before sending external emails, deploying to prod, moving money, agents stop and ask.

Full audit log

Every action taken, every tool call, every output rendered. Searchable forever.

Instant pause

One command pauses the agent. One disconnects it from a tool. No tickets, no waiting.

Per-channel scoping

Define what the agent can see and do, channel by channel. Default-deny posture.

Workspace isolation

Hard tenant boundary. Your data never crosses into another engagement. Period.

Data deletion

Full workspace wipe on request. Removes stored data, learned context, integration history.

07 / DEPLOY

Three steps. Two weeks. Live.

Bespoke · agency-style · no SaaS
01

Discover

One week with your team. Map the workflows that hurt. Quote the deploy.

02

Deploy

Bespoke agents in your Slack and WhatsApp. Tools connected. First workflows live.

03

Run

Monthly retainer. Tune. Add workflows. Widen the rift.

Send an agent
through the rift.

Thirty-minute call. We look at your ops, decide if there's a fit, and quote the discovery.