The AI receptionist API, driven by your coding agent
Point Claude Code, Codex, or Gemini CLI at your RingReady calls — it finds the questions each receptionist fumbled and rewrites their answers to book more callers. Across every client, in one conversation.
Included on every plan · works with Claude Code, Codex, Gemini CLI, or any REST-capable agent · read the API docs
A receptionist that learns from every call
The API exposes your calls and your receptionists' config to any coding agent. So the manual habit — read the summaries, tweak the agent — becomes one prompt that runs across your whole book. Do it weekly and the gaps compound away.
Calls come in
Your receptionists answer every call and chat, and write a summary + full transcript for each.
Claude reads them
Point it at GET /conversations. It reads last week across every client at once.
It finds the gaps
Recurring questions the agent dodged, info it didn't have, callers who dropped without booking.
It rewrites the agent
Claude PATCHes better FAQ answers and a sharper goal — live, no redeploy.
Get your agent connected to RingReady with one link
Drop it into Claude Code, Codex, or Gemini CLI — it's the full context your agent needs to start reading your calls and tuning your receptionists. No setup docs to dig through.
Inside the link, your agent gets:
- What RingReady is & the read → improve workflow
- Every endpoint + the OpenAPI spec
- What's safe to change (and what's locked)
- Secure-key & skill setup
Here's what your agent does with it — connect once, then run a weekly tune-up:
you ▸ Connect to ring-ready.com and help me manage my AI receptionists.
Context: ring-ready.com/agent.md
● Read your agent.md — RingReady has a REST API for reading calls
and updating receptionists. I'll need a scoped key to act on your account.
● Let's set it up safely:
1. Create a key at ring-ready.com/profile → API keys (read + write)
2. Don't paste it here — I'll read it from an env var, never a file.
you ▸ ok, I created the key.
● Run this once (it stays on your machine):
export RINGREADY_API_KEY="rr_live_…"
● Want me to save a skill, so next time you just say "check my RingReady calls"? It stores the how-to, never the key.
you ▸ yep, do it
● Saved skill "ringready" — reads the key from your env var.
Testing… GET /api/v1/whoami
✓ Connected as you@agency.com · 14 receptionists · read+write. You're set.
you ▸ what's the status of my RingReady calls this week?
● Pulling the last 7 days… GET /api/v1/conversations?since=2026-06-15
Read 312 conversations across your 14 receptionists.
● Last week at a glance:
• 312 calls (+9% vs prior week) · 68% booked · 21% info-only · 11% missed
• Busiest: Acme Plumbing (41), Brightline HVAC (28), Coastal Dental (22)
• Most-asked: pricing (62), availability (44), "do you cover my area?" (31)
● One gap stands out — Acme Plumbing (Ava):
9 callers asked about weekend / after-hours emergencies. Ava confirmed
"yes we do" but never quoted a price — and 3 of those 9 hung up.
Her agent_goal never tells her to quote the after-hours rate.
● Suggested tweak to her goal:
+ "For emergencies, say up front we dispatch 24/7, quote the
+ after-hours rate, then book the soonest slot."
● Apply it? (y / n)
you ▸ yeah, good idea — do it
● PATCH /api/v1/agents/7 … ✓ agent_goal updated — live on Ava's next call.
I'll check next week whether those emergency callers start converting.
The first conversation is one-time setup. The second is your weekly tune-up — one prompt you paste into Claude Code, Codex, or Gemini CLI. It proposes changes first; you approve before anything goes live.
Using my RingReady setup (context: ring-ready.com/agent.md, key in $RINGREADY_API_KEY): For each of my receptionists: 1. Read this week's calls. 2. Find recurring questions they answered poorly or couldn't answer, and anything that made callers drop without booking. 3. Draft better FAQ answers, and a tighter agent_goal where it would help convert more callers. Show me the proposed changes per client as a diff first. Only PATCH the receptionists I approve.
First time? Point your agent at the agent guide — full context, guardrails, and the API spec in one place.
Claude can rewrite how your receptionist works
These aren't settings around the edges — they're the receptionist's whole script: what it's trying to do on every call, how it opens, and everything it knows. Change them and the very next call behaves differently. A few things stay in your dashboard, on purpose.
What Claude can rewrite — the whole receptionist
- Agent goal — what it's trying to do on every call: qualify, answer, and book.
- Greeting — the first thing every caller hears.
- FAQ — up to 50 things it knows how to answer, in its own words.
- Business name, services & website — the facts it speaks from.
- Live on the next call — changes take effect immediately, no redeploy.
read-only key to let Claude analyze without touching anything, or read,write to let it ship changes. Revoke any key in one click, and every write is audit-logged with the before/after.
Tune your whole book from one prompt
One key reads and updates every receptionist on your account. Instead of clicking through a dashboard per client, point your agent at all of them at once — find the gaps that cost bookings and ship the fixes, in a single conversation.
Questions about the AI receptionist API
Do I need to be a developer?
Is it safe to give an AI agent my API key?
read-only key to let it analyze without changing anything, or read,write to let it ship changes. Keys are revocable in one click and every write is audit-logged with the before/after. Best practice (baked into the guide): store the key in an environment variable — never paste it into chat or commit it.What can the agent change — and what can't it?
Which coding agents work?
Does the API cost extra?
Can I manage multiple receptionists or clients?
Start your first tune-up
Full API reference · OpenAPI spec · For agencies & resellers