retellai-pack
Complete Retell AI integration skill pack with 30 skills covering AI voice agents, phone automation, conversational AI, and call center solutions. Flagship+ tier vendor pack.
Installation
Open Claude Code and run this command:
/plugin install retellai-pack@claude-code-plugins-plus
Use --global to install for all projects, or --project for current project only.
What It Does
> Claude Code skill pack for Retell AI — AI voice agents, phone call automation, LLM-powered conversations, and telephony integration (30 skills)
Skills (30) plugin-local skills
Retell AI advanced troubleshooting — AI voice agent and phone call automation.
Retell AI Advanced Troubleshooting
Overview
Implementation patterns for Retell AI advanced troubleshooting — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for advanced troubleshooting
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Isolate a voice-call failure without changing production routing
Reproduce the reported failure with a test number and a non-production agent version. Capture the call identifier, selected agent version, webhook response code, and timestamp; redact caller audio and personal data from the ticket. Compare that evidence with one successful test call before changing prompts, transfers, or routing. Promote the smallest verified correction through the preview path, then retain the before/after identifiers for rollback.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI architecture variants — AI voice agent and phone call automation.
Retell AI Architecture Variants
Overview
Implementation patterns for Retell AI architecture variants — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for architecture variants
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Choose a queue-backed architecture for bursty inbound calls
For a campaign that creates short spikes, keep the public number attached to a stable routing layer and place CRM enrichment behind an asynchronous queue. Start with a preview agent that records only synthetic test calls, set a clear timeout for the enrichment request, and configure a human-transfer fallback. Measure latency and transfer rate during the canary before selecting the variant for production traffic.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI ci integration — AI voice agent and phone call automation.
Retell AI Ci Integration
Overview
Implementation patterns for Retell AI ci integration — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for ci integration
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Validate an agent configuration in a pull request
In CI, lint the agent configuration and run a scripted call against a dedicated test agent using a masked test credential. Assert the expected tool invocation, handoff outcome, and redacted transcript markers, but never upload real call audio or customer data. Publish the test result as a PR artifact and require a separate promotion step before the tested configuration reaches a live number.
Resources
Next Steps
See related Retell AI skills for more workflows.
Diagnose and fix Retell AI voice agent errors: call failures, webhook issues, voice quality.
Retell AI Common Errors
Overview
Quick reference for the top Retell AI errors and their solutions.
Prerequisites
retell-sdkinstalled- API key configured
Instructions
Error 1: 401 Unauthorized
RetellError: 401 — Invalid API key
Fix: Verify API key in Retell Dashboard. Ensure RETELL_API_KEY starts with key_.
Error 2: Call Fails Immediately
RetellError: 400 — Invalid phone number format
Fix: Use E.164 format: +14155551234. Both from_number and to_number must be valid.
Error 3: Agent Not Responding
Call connected but agent says nothing
Fix: Check LLM configuration:
const llm = await retell.llm.retrieve(agent.response_engine.llm_id);
console.log(`Model: ${llm.model}`);
console.log(`Prompt length: ${llm.general_prompt.length} chars`);
// Ensure general_prompt is not empty and gives clear instructions
Error 4: Function Call Timeout
Function call to https://your-api.com/endpoint timed out
Fix: Your function endpoint must respond within 5 seconds. Offload heavy work:
app.post('/functions/lookup', async (req, res) => {
// Respond immediately with acknowledgment
const result = await quickLookup(req.body.args);
res.json({ result: `Found: ${result.name}` });
// Do NOT run async work before responding
});
Error 5: Webhook Not Receiving Events
No webhook events received after call
Fix: Set webhook_url on the agent, not just in Dashboard settings:
await retell.agent.update(agentId, {
webhook_url: 'https://your-app.com/webhooks/retell',
});
Error 6: Voice Quality Issues
Agent voice sounds robotic/choppy
Fix: Check network latency to Retell servers. Use a voice optimized for your use case. Try different voice IDs.
Output
- Error identified and root cause found
- Fix applied and verified
- Call successfully completed
Error Handling
| HTTP Code | Meaning | Retryable |
|---|---|---|
| 400 | Bad request | No — fix params |
| 401 | Invalid API key | No — fix key |
| 404 | Agent/call not found | No — fix ID |
| 429 | Rate limited | Yes — backoff |
| 500+ | Server error | Yes — retry |
Examples
Triage a failed outbound ca
Retell AI core workflow a — AI voice agent and phone call automation.
Retell AI Core Workflow A
Overview
Build and configure voice agents with custom prompts, function calling, and call flow logic.
Prerequisites
- Completed
retellai-hello-world
Instructions
Step 1: Agent with Function Calling
const llm = await retell.llm.create({
model: 'gpt-4o',
general_prompt: `You are a booking assistant for Dr. Smith's office.
- Help callers book, reschedule, or cancel appointments
- Collect: name, phone, preferred date/time
- Confirm all details before booking`,
functions: [
{
name: 'book_appointment',
description: 'Book a new appointment',
parameters: {
type: 'object',
properties: {
patient_name: { type: 'string' },
phone: { type: 'string' },
date: { type: 'string', description: 'YYYY-MM-DD format' },
time: { type: 'string', description: 'HH:MM format' },
},
required: ['patient_name', 'phone', 'date', 'time'],
},
url: 'https://your-api.com/appointments',
speak_during_execution: true,
speak_after_execution: true,
},
],
});
Step 2: Configure Voice and Behavior
const agent = await retell.agent.create({
response_engine: { type: 'retell-llm', llm_id: llm.llm_id },
voice_id: '11labs-Rachel',
agent_name: 'Dr. Smith Booking Agent',
language: 'en-US',
opt_out_sensitive_data_storage: false,
end_call_after_silence_ms: 10000, // End call after 10s silence
max_call_duration_ms: 300000, // 5-minute max
enable_backchannel: true, // "mhm", "yeah" responses
boosted_keywords: ['appointment', 'schedule', 'Dr. Smith'],
});
Step 3: Update Agent Configuration
await retell.agent.update(agent.agent_id, {
voice_id: '11labs-Dorothy', // Change voice
end_call_after_silence_ms: 15000,
});
Output
- Agent with custom LLM prompt and function calling
- Voice and behavior configuration
- Real-time function execution during calls
Error Handling
| Error | Cause | Solution |
|---|---|---|
| Function not triggering | Prompt doesn't guide to function | Include function use in prompt |
| Voice quality issues | Wrong voice selection | Test different voices |
| Call ending too early | Short silence timeout | Increase end_call_after_silence_ms |
Examples
Create and validate a new appointment-intake agent
B
Retell AI core workflow b — AI voice agent and phone call automation.
Retell AI Core Workflow B
Overview
Manage phone calls: outbound campaigns, call transfers, recordings, and concurrent call handling.
Prerequisites
- Completed
retellai-core-workflow-a
Instructions
Step 1: Outbound Call Campaign
const phoneNumbers = ['+14155551001', '+14155551002', '+14155551003'];
for (const number of phoneNumbers) {
try {
const call = await retell.call.createPhoneCall({
from_number: process.env.RETELL_PHONE_NUMBER!,
to_number: number,
override_agent_id: agentId,
metadata: { campaign: 'appointment-reminder', date: '2026-04-01' },
});
console.log(`Called ${number}: ${call.call_id}`);
} catch (err) {
console.error(`Failed to call ${number}: ${err.message}`);
}
// Rate limit: space calls apart
await new Promise(r => setTimeout(r, 2000));
}
Step 2: List and Filter Calls
const calls = await retell.call.list({
sort_order: 'descending',
limit: 20,
});
for (const call of calls) {
console.log(`${call.call_id}: ${call.call_status} — ${call.end_timestamp - call.start_timestamp}ms`);
}
Step 3: Get Call Recording and Transcript
const callDetail = await retell.call.retrieve(callId);
if (callDetail.recording_url) {
console.log(`Recording: ${callDetail.recording_url}`);
}
if (callDetail.transcript) {
console.log(`Transcript:\n${callDetail.transcript}`);
}
Output
- Outbound call campaign with rate limiting
- Call listing with status and duration
- Recordings and transcripts retrieved
Error Handling
| Error | Cause | Solution |
|---|---|---|
| Call fails immediately | Bad phone number format | Use E.164 format |
| No recording | Recording not enabled | Enable in agent settings |
| Concurrent limit | Too many active calls | Upgrade plan or queue calls |
Examples
Run a bounded outbound appointment reminder campaign
Upload only consented test contacts to a development campaign and set a small concurrency limit before enabling any production queue. For each test call, verify the E.164 number formatting, expected agent version, completion status, and redacted transcript retrieval. If a call result is ambiguous, query the call identifier before retrying so an operator does not create duplicate calls or overwrite the original outcome.
Resources
Next Steps
Handle call events:
Retell AI cost tuning — AI voice agent and phone call automation.
Retell AI Cost Tuning
Overview
Implementation patterns for Retell AI cost tuning — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for cost tuning
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Reduce cost without degrading a required human handoff
Measure average call duration, transfer rate, and failed-call retries for a representative non-production sample before changing prompt length or model settings. Trial the lower-cost configuration on a small canary, retaining the existing agent version as rollback. Keep the human-handoff trigger unchanged unless the service owner approves a new threshold; a shorter call is not a saving if it silently increases unresolved customer requests.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI data handling — AI voice agent and phone call automation.
Retell AI Data Handling
Overview
Implementation patterns for Retell AI data handling — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for data handling
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Minimize transcript data in a support escalation
For a failed call, retain the call identifier, timestamps, selected agent version, and error classification in the incident record. Do not paste full recordings or transcripts into a ticket; redact names, phone numbers, and payment details before attaching a narrow excerpt. Set the retention owner and expiry for the exported evidence, then delete the copy after the investigation and any agreed remediation are complete.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI debug bundle — AI voice agent and phone call automation.
Retell AI Debug Bundle
Overview
Implementation patterns for Retell AI debug bundle — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for debug bundle
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Assemble a safe support bundle for an intermittent failure
Collect the affected call ID, agent version, webhook status, API response headers, and deployment revision from a test or consented call. Exclude API keys, raw audio, and full transcripts from the archive; include a redacted timeline instead. Reproduce against a preview agent before changing the live configuration, and preserve the bundle checksum so the support handoff can be matched to the exact evidence reviewed.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI deploy integration — AI voice agent and phone call automation.
Retell AI Deploy Integration
Overview
Implementation patterns for Retell AI deploy integration — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for deploy integration
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Promote a tested agent configuration through a preview route
Create a versioned agent configuration and attach it to an internal preview number first. Exercise a synthetic call that verifies the expected webhook, tool permissions, and fallback transfer, then record the configuration version and test result. Promote only that reviewed version to the production routing rule, while retaining the previous version and its routing target for a fast, auditable rollback if call quality or completion rate regresses.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI enterprise rbac — AI voice agent and phone call automation.
Retell AI Enterprise Rbac
Overview
Implementation patterns for Retell AI enterprise rbac — voice agent and telephony platform.
Prerequisites
- Completed
retellai-install-authsetup
Instructions
Step 1: SDK Pattern
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
const agents = await retell.agent.list();
console.log(`Agents: ${agents.length}`);
Output
- Retell AI integration for enterprise rbac
Error Handling
| Error | Cause | Solution |
|---|---|---|
| 401 Unauthorized | Invalid API key | Check RETELL_API_KEY |
| 429 Rate Limited | Too many requests | Implement backoff |
| 400 Bad Request | Invalid parameters | Check API documentation |
Examples
Grant a release operator access without granting call-data access
Create a role limited to publishing an approved agent version and inspecting deployment status. Do not include transcript, recording, billing, or user management permissions in that role. Test the role against a preview agent: it should promote the designated version but fail to retrieve protected call content. Record the role, environment, and approver before assigning it to a new operator, and remove the assignment when the release window closes.
Resources
Next Steps
See related Retell AI skills for more workflows.
Retell AI hello world — AI voice agent and phone call automation.
Retell AI Hello World
Overview
Create your first Retell AI voice agent and make a test phone call.
Prerequisites
- Completed
retellai-install-auth - A phone number registered in Retell AI Dashboard (or use web call for testing)
Instructions
Step 1: Create an LLM Configuration
import Retell from 'retell-sdk';
const retell = new Retell({ apiKey: process.env.RETELL_API_KEY! });
// Create LLM configuration (what the agent says)
const llm = await retell.llm.create({
model: 'gpt-4o',
general_prompt: `You are a friendly receptionist for Acme Corp.
- Greet callers warmly
- Ask how you can help
- Take messages if needed
- Be concise and professional`,
});
console.log(`LLM created: ${llm.llm_id}`);
Step 2: Create a Voice Agent
const agent = await retell.agent.create({
response_engine: {
type: 'retell-llm',
llm_id: llm.llm_id,
},
voice_id: '11labs-Adrian', // Choose from available voices
agent_name: 'Acme Receptionist',
});
console.log(`Agent created: ${agent.agent_id}`);
Step 3: Make a Test Phone Call
// Outbound call (requires a registered phone number)
const call = await retell.call.createPhoneCall({
from_number: '+14155551234', // Your Retell number
to_number: '+14155555678', // Destination
override_agent_id: agent.agent_id,
});
console.log(`Call initiated: ${call.call_id}`);
Step 4: Or Test with Web Call
// Web call (no phone number needed — great for testing)
const webCall = await retell.call.createWebCall({
agent_id: agent.agent_id,
});
console.log(`Web call URL: ${webCall.call_id}`);
// Use retell-client-js-sdk to connect from browser
Step 5: Check Call Status
const callDetail = await retell.call.retrieve(call.call_id);
console.log(`Status: ${callDetail.call_status}`);
console.log(`Duration: ${callDetail.end_timestamp - callDetail.start_timestamp}ms`);
if (callDetail.transcript) {
console.log(`Transcript: ${callDetail.transcript}`);
}
Output
- LLM configuration with custom prompt
- Voice agent with selected voice
- Test call initiated (phone or web)
- Call status and transcript retrieved
Error Handling
| Error | Cause | Solution | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
422 Invalid voice_id |
Unknown voice | List available voices in Dashboard | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
400 No phone number |
Number not registered | Register number in Dashboard first | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Call not connecting | Destination unreachable | Try web call for test
retellai-incident-runbook
View full skill →
Retell AI incident runbook — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Incident RunbookOverviewImplementation patterns for Retell AI incident runbook — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesContain a spike in failed transfersWhen transfer failures cross the alert threshold, freeze new agent promotions and route new calls to the last known-good agent or human queue. Capture call identifiers, agent versions, transfer destinations, and response codes without copying recordings into the incident channel. Compare a synthetic preview call with the affected production version, make one reversible routing change, and document the rollback condition before restoring normal traffic. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-install-auth
View full skill →
Retell AI install auth — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Install AuthOverviewInstall the Retell AI SDK and configure API key authentication for building voice agents. Prerequisites
InstructionsStep 1: Install SDK
Step 2: Configure Environment
Step 3: Initialize Client (Node.js)
Step 4: Initialize Client (Python)
Output
Error Handling
ExamplesConfigure a least-privilege development credentialStore a non-production API key in the local secret manager or CI secret store, not in source control or a command history. Verify it by listing development agents and record only the account, environment, and verification time. If the test returns an authorization error, revoke the invalid credential rather than trying alternate keys in logs; use a separate production credential only in a protected deployment environment. ResourcesNext StepsCreate your first agent:
retellai-known-pitfalls
View full skill →
Retell AI known pitfalls — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Known PitfallsOverviewImplementation patterns for Retell AI known pitfalls — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesAvoid a duplicate-call retry after a network timeoutAn outbound request times out after submission. Before retrying, look up the request correlation or call identifier to determine whether Retell already created the call. If state is unknown, hold the retry and escalate with the redacted request metadata; sending a second request can contact the same person twice. Add bounded retries only where the operation is explicitly idempotent and monitor the duplicate-call rate after the change. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-load-scale
View full skill →
Retell AI load scale — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Load ScaleOverviewImplementation patterns for Retell AI load scale — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesLoad-test an inbound queue without contacting real callersUse a preview number and synthetic caller identities to ramp concurrency in small steps. Record queue time, model latency, transfer rate, error rate, and the configured concurrency limit at each step. Stop the test when the agreed latency or error budget is crossed instead of compensating with unbounded retries. Use the result to set an initial production ceiling and retain the last known-good limit as the rollback value. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-local-dev-loop
View full skill →
Retell AI local dev loop — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Local Dev LoopOverviewImplementation patterns for Retell AI local dev loop — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesIterate on a prompt with a deterministic local test setKeep a small suite of synthetic caller intents, including a successful request, an unclear request, and a required human handoff. Run that suite against a development agent after each prompt or tool-schema change and compare only redacted outcomes such as route, tool arguments, and completion state. Do not point local development at a production phone number; publish a versioned preview only after the suite continues to meet its expected behavior. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-migration-deep-dive
View full skill →
Retell AI migration deep dive — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Migration Deep DiveOverviewImplementation patterns for Retell AI migration deep dive — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesMigrate one agent while preserving a rollback pathExport the current agent configuration and record its version, routing rule, and supported handoff behavior before making schema or prompt changes. Import the candidate into a preview environment, replay synthetic scenarios, and compare the redacted outcomes with the old version. Move a small canary route only after those checks pass; leave the previous configuration intact until the canary completion and escalation rates remain within the agreed bounds. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-multi-env-setup
View full skill →
Retell AI multi env setup — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Multi Env SetupOverviewImplementation patterns for Retell AI multi env setup — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesKeep development and production credentials separatedCreate distinct Retell projects or explicitly named agent groups for development, staging, and production. Each environment receives its own secret reference and phone routing configuration, so a local test cannot accidentally create a live call. Verify the separation by listing agents with the development credential and confirming that production identifiers are absent; promote configuration through a controlled CI environment instead of copying keys between files. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-observability
View full skill →
Retell AI observability — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI ObservabilityOverviewImplementation patterns for Retell AI observability — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesAlert on a regression while protecting call contentEmit structured metrics for call completion, transfer outcome, latency, and provider errors using agent version and environment as bounded labels. When an alert fires, link the affected call identifiers to a restricted investigation record instead of attaching transcripts to the dashboard. Compare a synthetic call against the last known-good version, then make one reversible routing or configuration change and watch the same metrics through the defined recovery window. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-performance-tuning
View full skill →
Retell AI performance tuning — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Performance TuningOverviewImplementation patterns for Retell AI performance tuning — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesImprove response latency with a measured canaryCapture baseline end-to-end latency, model latency, transfer outcome, and completion rate for a synthetic test suite before changing an agent setting. Apply one prompt, model, or integration change to a preview version and repeat the same suite. Promote to a small canary only if latency improves without a drop in completion or an increase in handoffs; otherwise restore the prior version and retain the comparison as the performance decision record. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-policy-guardrails
View full skill →
Retell AI policy guardrails — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Policy GuardrailsOverviewImplementation patterns for Retell AI policy guardrails — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesRestrict a booking agent to an approved action boundaryGive an appointment agent only the lookup and create-booking tools needed for its flow, with explicit argument validation and a human handoff for refunds, medical questions, or account changes. Test synthetic prompts that attempt each forbidden action and verify that the agent refuses or transfers rather than improvising. Version the policy with the agent and review deviations from the guardrail before widening any permission. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-prod-checklist
View full skill →
Retell AI prod checklist — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Prod ChecklistOverviewImplementation patterns for Retell AI prod checklist — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesApprove a limited production launchBefore routing a live number, verify the exact agent version, environment secret references, consent and disclosure text, fallback transfer destination, and on-call owner. Run a final synthetic call through the production-like preview route and record the result without retaining audio. Start with a small traffic percentage and a defined rollback trigger, then expand only when completion, escalation, and error metrics remain within the release bounds. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-rate-limits
View full skill →
Retell AI rate limits — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Rate LimitsOverviewImplementation patterns for Retell AI rate limits — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesHandle a burst without retrying calls indefinitelyWhen the API returns a rate-limit response, place only the idempotent request metadata in a bounded queue and honor the provider retry guidance before the next attempt. Track queue age, retry count, and error class; do not retry an ambiguous call-creation request until its call identifier is checked. Alert an operator when the queue age exceeds the service target, and shed nonessential work rather than allowing retries to amplify the burst. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-reference-architecture
View full skill →
Retell AI reference architecture — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Reference ArchitectureOverviewImplementation patterns for Retell AI reference architecture — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesSeparate the call path from slow business-system enrichmentRoute the voice agent through a stable API boundary that validates inputs and returns a bounded response, while sending slow CRM or analytics enrichment to an asynchronous worker. The agent receives a clear timeout and human-transfer fallback if the business system is unavailable. Exercise this topology with a synthetic outage before launch, then document which component owns retries, audit events, and the rollback to the prior routing configuration. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-reliability-patterns
View full skill →
Retell AI reliability patterns — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Reliability PatternsOverviewImplementation patterns for Retell AI reliability patterns — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesDesign a safe fallback for a dependent CRM outageGive the voice flow a short, explicit timeout for CRM lookup and a fallback response that either schedules a follow-up or transfers to a human queue. Test the fallback by making the preview CRM endpoint return a controlled error, then confirm no call is retried or double-booked. Track the fallback rate separately from general call failures and remove the temporary routing change only after the dependency and a synthetic recovery call both succeed. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-sdk-patterns
View full skill →
Production-ready Retell AI SDK patterns for voice agent applications.
ReadWriteEdit
Retell AI SDK PatternsOverviewProduction-ready patterns for Retell AI: client singletons, typed agent configurations, call management, and error handling. Prerequisites
InstructionsStep 1: Singleton Client
Step 2: Typed Agent Configuration
Step 3: Call Manager with Retry
Step 4: Batch Call Campaign
retellai-security-basics
View full skill →
Retell AI security basics — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Security BasicsOverviewImplementation patterns for Retell AI security basics — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesRotate an exposed development credential safelyDisable the exposed development key in the provider console, create a scoped replacement, and update only the development secret reference. Verify the new key with a harmless agent-list request while ensuring command output does not print its value. Review access logs and affected preview configurations, then document the incident with the credential identifier and rotation time rather than copying key material into the ticket or source repository. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-upgrade-migration
View full skill →
Retell AI upgrade migration — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Upgrade MigrationOverviewImplementation patterns for Retell AI upgrade migration — voice agent and telephony platform. Prerequisites
InstructionsStep 1: SDK Pattern
Output
Error Handling
ExamplesUpgrade an SDK through a compatibility canaryPin the candidate SDK version in a branch and run the development agent suite against synthetic create-call, retrieval, error, and webhook scenarios. Record the old and new package versions and any changed request shapes, then deploy a preview configuration before altering production dependencies. Keep the prior lockfile and agent version available until the canary has met its error and latency limits, so rollback does not require reconstructing the old runtime. ResourcesNext StepsSee related Retell AI skills for more workflows.
retellai-webhooks-events
View full skill →
Retell AI webhooks events — AI voice agent and phone call automation.
ReadWriteEditBash(npm:*)Bash(curl:*)Grep
Retell AI Webhooks EventsOverviewHandle Retell AI webhook events for call lifecycle, transcripts, and function execution. Prerequisites
InstructionsStep 1: Configure Webhook URL
Step 2: Webhook Endpoint
Step 3: Handle Function Calls During Conversation
Output
Error Handling
ExamplesProcess a duplicate webhook delivery safelyVerify the webhook signature, derive a stable event identifier, and record the event before invoking any downstream side effect. When the same identifier is received again, acknowledge it without re-sending a notification, creat Ready to use retellai-pack?Related Pluginssupabase-packComplete Supabase integration skill pack with 30 skills covering authentication, database, storage, realtime, edge functions, and production operations. Flagship+ tier vendor pack. /plugin install supabase-pack@claude-code-plugins-plus
vercel-packComplete Vercel integration skill pack with 30 skills covering deployments, edge functions, preview environments, performance optimization, and production operations. Flagship+ tier vendor pack. /plugin install vercel-pack@claude-code-plugins-plus
clay-packComplete Clay integration skill pack with 30 skills covering data enrichment, waterfall workflows, AI agents, and GTM automation. Flagship+ tier vendor pack. /plugin install clay-pack@claude-code-plugins-plus
cursor-packComplete Cursor integration skill pack with 30 skills covering AI code editing, composer workflows, codebase indexing, and productivity features. Flagship+ tier vendor pack. /plugin install cursor-pack@claude-code-plugins-plus
exa-packComplete Exa integration skill pack with 30 skills covering neural search, semantic retrieval, web search API, and AI-powered discovery. Flagship+ tier vendor pack. /plugin install exa-pack@claude-code-plugins-plus
firecrawl-packComplete Firecrawl integration skill pack with 30 skills covering web scraping, crawling, markdown conversion, and LLM-ready data extraction. Flagship+ tier vendor pack. /plugin install firecrawl-pack@claude-code-plugins-plus
Tags
retellairetellvoice-aiphone-agentsconversational-aicall-centerivr
|