Aller au contenu
TakoAPI
Retour aux skills

telnyx-rag

par teamtelnyx

Web & Frontend Development
647Téléchargements rapportés par clawskills.sh
0Vues
0 mention J'aimeConnecte-toi pour aimer ce skill
—Pas encore noté
Ta note

Résumé

Semantic search and Q&A over workspace files using Telnyx Storage + AI embeddings.

Description

Ce que fait ce skill

Indexes workspace files into Telnyx Storage with AI-generated embeddings, enabling semantic search and RAG-powered Q&A over memory, knowledge, and skills directories. Files are chunked by document type, synced incrementally, and retrieved with multi-signal reranking before LLM generation.

Uses Telnyx's native embedding, similarity search, and inference APIs in one pipeline, removing the need to wire together a separate vector database, embedding service, and LLM provider.

Quand l'utiliser

  • Asking questions about past meeting decisions stored in knowledge files
  • Searching personal notes and preferences by meaning, not keyword
  • Querying documentation for deployment or onboarding steps
  • Auto-syncing workspace memory on a cron schedule for always-current retrieval
  • Retrieving project history from indexed Slack message exports

Exemple de workflow

ENTRÉE

User asks: What is the porting process?

AGENT
  1. 1

    Runs ./sync.py to upload changed workspace files to Telnyx Storage with smart chunking

  2. 2

    Triggers ./sync.py --embed to generate vector embeddings for uploaded chunks

  3. 3

    Calls ./ask.py with the query, retrieving the top 20 candidate chunks

  4. 4

    Reranks chunks by semantic similarity, keyword overlap, and memory priority

  5. 5

    Calls Telnyx Inference LLM with the top 8 context chunks and returns the answer with source citations

SORTIE

Natural language answer identifying the porting process steps, with references to the source files

Prérequis

  • TELNYX_API_KEY environment variable
  • Telnyx account to generate the API key at portal.telnyx.com

Installation

clawhub install teamtelnyx/telnyx-rag

Exécute la commande CLI, ou colle l'instruction dans Claude Code, Cursor, Codex ou l'agent de code que tu utilises.

Liens