rg://email-agent

Email replies,
grounded in your docs.

Orepli reads every inbound email, finds the answer in your knowledge base, and drafts a reply. Auto-send when it's confident. Escalate when it isn't. Every step cited and logged.

Gmail OAuthIMAP / SMTPClaude 3.5pgvector
The problem

Your team answers the same ten questions a hundred times a day.

Support inboxes are a queue of repeats. The answer is already in your docs — a policy, a product spec, a how-to. But the human in the loop still has to find it, phrase it, and paste it. Meanwhile first-response time slips, tone drifts, and nobody can audit why a reply said what it said.

  • Repetitive questions eat senior time. The answer is never the hard part — the lookup is.
  • Docs update. Replies don't. Last month's policy ships this month's email.
  • You have no audit trail. Just a Gmail thread and a memory.
  • Escalations happen by vibe, not by rule.
How it works

From inbound to sent in under ten seconds.

Every email runs through four gates: connect, ground, classify, reply. You keep control at each one. Turn auto-send on when you're comfortable. Leave it off forever if you prefer every reply reviewed.

01

Connect your inbox

Plug in Gmail in a click, or any IMAP/SMTP mailbox with encrypted credentials. Multiple mailboxes per workspace — one per queue, per team, per region.

Gmail OAuth · IMAP/SMTP
02

Upload your knowledge

Drop in FAQs, help articles, refund policies, product specs — anything in plain text or Markdown. Orepli chunks it, embeds it, and scopes it to the mailbox that should use it.

TXT · MD · Policies · SOPs
03

Tune the rules

Pick a tone, set the confidence threshold that has to be crossed before auto-send, and choose how many KB chunks to retrieve per reply. Start in draft-only mode until you trust the output.

Confidence threshold · Tone · Top-K
04

Watch it work

Every inbound email runs through the pipeline. You see the full timeline — classification, retrieved snippets with citations, the draft, and either the send confirmation or the escalation reason.

Classified · Cited · Drafted · Sent or escalated
What's inside

Opinionated defaults. Every knob where you need one.

Built on the stack that serious RAG teams already run — Postgres with pgvector, Claude for generation, explicit citations, per-tenant isolation. No black box.

01

Grounded retrieval

Every draft cites the KB chunks it was built from. pgvector under the hood. Swap the embedder if you need HF, OpenAI, or local.

02

Confidence gating

Replies only auto-send when model confidence clears your threshold. Everything else becomes a draft or a routed escalation.

03

Tone control

Per-mailbox voice — terse, warm, formal, technical. Different queues, different voices, same knowledge base.

04

Smart escalation

Edge cases forward to the right human with the reason, the retrieved context, and the draft we would have sent.

05

Multi-tenant audit

Received → classified → retrieved → drafted → sent. Every step is logged, queryable, and isolated per tenant.

06

Dead-letter queue

Failures never silently disappear. Retry, inspect, or escalate. Your ops team gets a real picture of what broke and why.

Proof trail

Every reply shows its work.

Cited chunks, confidence score, model, latency, token count. When a draft ships, you can trace back why — and when it doesn't, you see exactly which rule caught it.

TRACE_01Email #4821
intentrefund
confidence0.92 > 0.80 threshold
citedrefund-policy.md · shipping-sop.md
modelclaude-3-5-sonnet
tokens1,104 in · 226 out
latency9.34s
actionsent
escalatedno
Pricing

Usage-honest pricing. No per-seat nonsense.

Pick the plan that fits your mailbox count and monthly token budget. Overage is linear — no surprise bills, no throttles. Upgrade or downgrade any time.

Starter
$49 /mo

For a founder or a small team automating a single support queue.

  • 1 mailbox
  • 10M tokens / month
  • Gmail + IMAP/SMTP
  • Draft-only or auto-send
  • Email support
Start Starter
Scale
$199 /mo

For teams scaling RAG-native email across brands or regions.

  • 5 mailboxes
  • 100M tokens / month
  • Everything in Growth
  • Extra mailboxes $12/ea
  • Slack support
Start Scale
FAQ

Answers before you ask.

Still have questions? sales@orepli.com — we reply with Orepli, obviously.

Will Orepli send emails without my approval?

Only when the draft's confidence score is above the threshold you set on the mailbox. Everything under the threshold becomes a reviewable draft or gets escalated. You can also run a mailbox in draft-only mode indefinitely.

Which email providers work?

Gmail via OAuth, or any IMAP/SMTP inbox with TLS. Credentials are encrypted at rest with a tenant-scoped key. OAuth tokens are refreshed automatically.

What AI model does it use?

Claude from Anthropic by default, with a mock mode for local development. Embeddings default to a deterministic local hash embedder; you can switch to Hugging Face sentence-transformers or a managed provider via a single env var.

How do you handle my data?

Knowledge base documents are scoped to the mailbox you upload them to. Email content is stored per tenant in Postgres with pgvector. You can request export or deletion at any time. See the Privacy page for the full picture.

Can I try it before paying?

Yes. Register a workspace, connect a mailbox, run inbound emails through the simulator, and preview drafts before turning on auto-send. No card needed to evaluate.

Do you offer self-hosting?

Yes, for Scale customers and above. The stack is FastAPI + Postgres + pgvector + Next.js, deployable via Docker Compose. Contact sales@orepli.com to discuss.

Get started

Stop answering the same email. Start reviewing replies.

Spin up a workspace, connect a mailbox, upload a KB, and watch the pipeline work on your real inbox. No credit card to evaluate.