use-with-pipecat
Skill0 starsChanged 57 days agoMITView on GitHub
Connect Pipecat to the ThunderPhone Realtime endpoint. Use when someone asks for pipecat-thunderphone, ThunderPhoneRealtimeLLMService, an OpenAI Realtime-compatible Pipecat service, package installation, saved agent_id configuration, or Pipecat voice-session troubleshooting.
- Reads credentials
- Installs packages
- Sends data out
In this repository21 other files it gives its agents.
- AGENTS.mdAGENTS.md
- add-knowledgeSkill
- build-tool-integrationSkill
- create-agentSkill
- embed-web-widgetSkill
- get-phone-numberSkill
- handle-inbound-callsSkill
- iterate-on-agentSkill
- migrate-from-blandSkill
- migrate-from-elevenlabsSkill
- migrate-from-retellSkill
- migrate-from-vapiSkill
- place-outbound-callsSkill
- run-campaignSkill
- setup-api-keySkill
- setup-webhooksSkill
- test-agentSkill
- thunderphone-mcp-setupSkill
- thunderphone-prompt-builderSkill
- use-with-livekitSkill
- use-with-openai-realtimeSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 5 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Install the documented adapter: Use pip install pipecat-thunderphone in a clean environment.
- Distinctive: line 16
Select agent mode: Prefer a saved deployed ThunderPhone agent for shared prompt, voice, tools, knowledge, and testing.
- Very specific: line 21
Create the service: Import ThunderPhoneRealtimeLLMService from the installed adapter and pass the API key plus the documented saved …
- Distinctive: line 64
Verify end to end: Test connection, two-way audio, interruptions, tool calls, language switching, teardown, and error recovery.
- Distinctive: line 79
Follow use with Pipecat and the Realtime API.
- Distinctive: line 80
Keep the API key server-side and redact event payloads before logging.
- Distinctive: line 81
Do not claim latency or audio quality without a real audio run.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Use ThunderPhone with Pipecat
- Install the documented adapter. Use
pip install pipecat-thunderphonein a clean environment. Confirm the resolvedpipecat-aiversion is 1.8 or newer; do not mix incompatible transport examples from older releases. - Select agent mode. Prefer a saved deployed ThunderPhone agent for shared prompt, voice, tools, knowledge, and testing. Use inline session config only when the application must own those fields.
- Store the credential as
THUNDERPHONE_API_KEY. Never embed it in client JavaScript, recordings, or pipeline logs. - Create the service. Import
ThunderPhoneRealtimeLLMServicefrom the installed adapter and pass the API key plus the documented savedagent_idor inline session options. Wire the service between the application's chosen input/output transports as shown in the Pipecat guide. - Use the OpenAI Realtime protocol at
wss://api.thunderphone.com/v1/realtime. Let the adapter manage event and audio framing; do not combine REST call fields with realtime session fields. - Inspect server-side sessions during backend diagnosis with
POST /v1/realtime/sessionsonly when the integration requires explicit session creation.
{
"agent_id": 12
}curl --fail-with-body -X POST https://api.thunderphone.com/v1/realtime/sessions \
-H "Authorization: Bearer $THUNDERPHONE_API_KEY" \
-H "Content-Type: application/json" \
-H "X-ThunderPhone-Client: skills/use-with-pipecat@1.0" \
--data-binary @realtime-session.jsonimport json, os, requests
with open("realtime-session.json", encoding="utf-8") as f: payload = json.load(f)
r = requests.post("https://api.thunderphone.com/v1/realtime/sessions", json=payload,
headers={"Authorization": f"Bearer {os.environ['THUNDERPHONE_API_KEY']}",
"X-ThunderPhone-Client": "skills/use-with-pipecat@1.0"}, timeout=30)
r.raise_for_status(); print(r.json())import { readFile } from "node:fs/promises";
const r = await fetch("https://api.thunderphone.com/v1/realtime/sessions", { method: "POST",
headers: { Authorization: `Bearer ${process.env.THUNDERPHONE_API_KEY}`,
"Content-Type": "application/json", "X-ThunderPhone-Client": "skills/use-with-pipecat@1.0" },
body: await readFile("realtime-session.json", "utf8") });
if (!r.ok) throw new Error(await r.text()); console.log(await r.json());- Verify end to end. Test connection, two-way audio, interruptions, tool calls, language switching, teardown, and error recovery. Correlate client events with ThunderPhone call/session records.
Errors
| Symptom | Cause | Fix |
| Import fails | Adapter or compatible Pipecat version is missing | Install pipecat-thunderphone and verify the environment's dependency versions. |
| WebSocket returns unauthorized | Key is missing/invalid or auth is sent in the wrong place | Let the adapter use THUNDERPHONE_API_KEY; do not expose it to browser code. |
| Audio connects but behavior is stale | Saved agent draft was not deployed or wrong agent ID is used | Read the agent, deploy the intended revision, and reconnect. |
| Duplicate audio or events | Two services/transports own the same stream | Keep one audio path and one lifecycle owner per session. |
Source and safety rules
- Follow use with Pipecat and the Realtime API.
- Keep the API key server-side and redact event payloads before logging.
- Do not claim latency or audio quality without a real audio run.
- Pin compatible packages for production and retest before upgrades.
See references/pipecat-checklist.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

