didww/ai / skills/didww-connect
didww-connect
VerifiedConnect a DIDWW phone number to ANY AI voice platform: helps choose the platform, then wires the SIP trunk pair and number. Use whenever the user wants a voice AI agent on a DIDWW number and has not committed to a platform, asks which AI platform to use with DIDWW, or names a platform that has no dedicated didww-to-X skill installed. Covers ElevenLabs, Vapi, Retell, Ultravox, Vogent, OpenAI Realtime, xAI, LiveKit and the enterprise-gated platforms.
| File | Kind | Size |
|---|---|---|
| .npmignore | other | 66 B |
| LICENSE | other | 1.1 KB |
| README.md | markdown | 2.2 KB |
| SKILL.md | skill md | 7.7 KB |
| bin/install.jsexecutable | script | 5.3 KB |
| install.shexecutable | script | 1.0 KB |
| package.json | manifest | 888 B |
| publish.shexecutable | script | 604 B |
| references/panels-and-billing.md | markdown | 6.2 KB |
| scripts/monitor.py | script | 5.7 KB |
| test.shexecutable | script | 488 B |
| tests/mock_didww.py | script | 3.0 KB |
| tests/test_monitor.py | script | 3.8 KB |
references/panels-and-billing.md
View on GitHub# Panels, Keys and Money: All Providers
Where things live in each provider's panel and how each one bills,
verified August 2026. Menu labels drift; treat paths as landmarks and
adapt if the user reports different screens. Always prefer MCP/OAuth over
keys where a column says so.
## DIDWW (the constant)
- Panel `my.didww.com` (signup `/users/sign_up`; sandbox panel
`my-sandbox.didww.com`).
- Automation: MCP `https://api.didww.com/mcp` with OAuth: NO key needed.
API key (script fallbacks only): panel menu API, then DIDWW API 3,
Create new API Key.
- Money: pure PREPAID. Billing, then Payment Methods: card, PayPal, wire;
Instant Payments for one-click top-up; Auto-charge toggle per method.
Low Balance Notification lives on the same page and is OFF by default:
switch it on with a threshold, every time, because services and number
renewals stop at zero balance.
- Outbound trunks: Voice, Outbound Trunks. NOT in the MCP; the user
creates them in the panel. A Get Access button there means the account
is not approved for outbound yet (application needs a business account
with company name and VAT/TAX ID; DIDWW answers within 48 business
hours). The per-trunk 24 Hour Limit (USD) defaults to 3000, minimum
50, and disables the trunk when reached; Capacity limit and a usage
notification at 80 percent sit next to it under Termination.
Credentials & IP-Based authentication needs at least one Allowed SIP
IP. Docs: `doc.didww.com/voice/outbound-trunks/`.
- Inbound trunks (MCP): `capacity_limit` caps concurrent calls per trunk.
## ElevenLabs
- Panel `elevenlabs.io/app`. Keys: `elevenlabs.io/app/settings/api-keys`
(workspace Service Account keys: profile icon, Workspace settings,
Service Accounts). Prefer the OAuth MCP
(`api.elevenlabs.io/v1/mcp`) for everything it covers.
- Money: monthly SUBSCRIPTION tiers granting credits (Free includes 15
agent minutes; Starter 6 USD/75 min; Creator 22 USD/275 min; Pro 99
USD/1,238 min). Extra agent minutes about 0.08 USD; concurrency above
the tier cap bills 2x. LLM usage and telephony are billed ON TOP of
the per-minute rate.
- Gotcha: agent minutes drain the same credit pool as TTS, so a chatty
agent eats the subscription fast; watch usage in the first week.
## Vapi
- Panel `dashboard.vapi.ai`. Keys: `/org/api-keys`, shown ONCE at
creation. Their MCP (`mcp.vapi.ai/mcp`, Bearer key) reads but cannot
create trunks/numbers.
- Money: PREPAID credits, 1 credit = 1 USD, card with auto-reload
(minimum 10 USD). 10 USD free credits, no card needed to start.
Platform 0.05 USD/min plus provider passthrough (real total usually
0.15 to 0.30 USD/min); numbers 2 USD/month.
- Gotcha: a NEGATIVE balance freezes the wallet, and if frozen at billing
time add-ons INCLUDING PHONE NUMBERS auto-cancel. Set auto-reload on
day one.
## Retell
- Panel `dashboard.retellai.com`. Keys: API Keys tab. Their MCP
(`mcp.retellai.com`, Bearer key) covers agents AND number import:
fullest automation of the group.
- Money: PREPAID credits with optional auto-recharge, Billing tab. 10 USD
free trial credits. Real cost roughly 0.07 to 0.31 USD/min; using your
own DIDWW trunk zeroes Retell's 0.015 USD/min telephony line, a real
saving worth telling the user.
- Gotcha: subscription items (numbers, extra knowledge bases,
concurrency) bill the card separately at cycle end, and calls
hard-block at zero credits.
## Ultravox
- Panel `app.ultravox.ai`. Keys: Settings, API Keys, Generate New Key.
No MCP.
- Money: usage-based, 0.05 USD/min agent time; SIP minutes are a
SEPARATE line (about 0.005 USD/min). 30 free minutes. Pay-as-you-go
has concurrency caps; Pro at 100 USD/month removes them.
- Gotcha: two meters run per call (agent time plus SIP time); quote both.
## Vogent
- Panel `app.vogent.ai`. Keys: API Keys page in the sidebar. No MCP.
- Money: per-minute, around 0.09 USD/min per their marketing, but there
is NO public pricing page; verify actual rates inside the panel before
the user scales anything.
## OpenAI (Realtime SIP)
- Panel `platform.openai.com`. Keys: `/api-keys`. Webhooks:
`/settings/project/webhooks` (create, copy the signing secret).
- Money: PREPAID credits, Settings then Billing, minimum purchase 5 USD.
Auto-recharge defaults to ON during setup; credits EXPIRE after one
year, non-refundable. Realtime bills in audio tokens, not minutes
(about 600 tokens/min heard, 1,200/min spoken); a typical phone
conversation lands around 0.05 to 0.10 USD/min of model cost depending
on talk ratio.
- Gotcha: the expiry and default auto-recharge surprise people; say both
out loud during setup.
## xAI (Grok voice)
- Panel `console.x.ai`. Keys: API Keys page. Billing: Billing, API spend
management (buy credits), Billing details for the card.
- Money: PREPAID credits, minimum top-up 25 USD, no free credits
documented; calls hard-fail the instant credits hit zero unless auto
top-up is configured (threshold, amount, monthly cap).
## LiveKit (Cloud or self-hosted)
- Panel `cloud.livekit.io`: project, Settings, API keys (API key plus a
secret shown ONCE); the Project URL and the SIP URI (a different
subdomain, never derived from the project URL) are on the same
settings page. No MCP; the SIP API is a
Twirp HTTP service, covered by the skill's `scripts/connect.py`.
- Money: Build tier FREE with no card (1,000 agent-session minutes,
1,000 SIP minutes, 5 concurrent sessions); Ship 50 USD/month (5,000
minutes, 20 concurrent); Scale 500 USD/month (50,000 minutes, up to
600 concurrent). Overage: agent session 0.01 USD/min, SIP 0.004
USD/min (0.003 on Scale). The STT, LLM and TTS the worker uses bill
separately at their own providers, or through LiveKit Inference
credits.
- Gotcha: three meters run per call (agent session, SIP minute, the
model providers), and a stopped worker means unanswered calls, not
an error message; monitor the worker, not only the balances.
## The advice that applies everywhere
Turn on every low-balance or auto-recharge safety the provider offers at
setup time, not after the first dead call; tell the user the realistic
per-minute all-in cost for their chosen stack before the first real
traffic; and remind them the DIDWW side is prepaid too, so two balances
need watching: that is what the hub's `scripts/monitor.py` checks.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |