Text-to-vector embeddings and semantic search using Telnyx AI. Generate embedding vectors via an OpenAI-compatible API — no OpenAI or Google API keys required.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Text-to-vector embeddings and semantic search using Telnyx AI. Generate embedding vectors via an OpenAI-compatible API — no OpenAI or Google API keys required.
Generate embedding vectors from text using Telnyx's OpenAI-compatible AI API. Convert any text to high-dimensional vectors for similarity comparisons, clustering, classification, or building custom search indexes — all with just a TELNYX_API_KEY. No OpenAI or Google API keys required.
Requirements
Python 3.8+ — stdlib only, no external dependencies
That's it. No pip install, no setup wizard, no external provider keys.
Text-to-Vector Embedding
Generate embedding vectors for any text input. The API is OpenAI-compatible, so existing integrations work out of the box.
Basic Usage
# Embed text (uses thenlper/gte-large by default)
./embed.py "text to embed"# Use a specific model
./embed.py "text to embed" --model intfloat/multilingual-e5-large
./embed.py --file input.txt
| ./embed.py --stdin
./embed.py --json
./embed.py --list-models
# Read from file
# Pipe from stdin
echo
"text to embed"
# JSON output (for scripting)
"text"
# List available models
Available Models
Model
Description
thenlper/gte-large
General text embeddings (default)
intfloat/multilingual-e5-large
Multilingual text embeddings
OpenAI-Compatible Client
The embeddings API is OpenAI-compatible, so you can use the OpenAI Python SDK with base_url pointed at Telnyx:
from embed import embed_text
result = embed_text("your text here")
for item in result.get("data", []):
vector = item["embedding"] # list of floats
dims = item["dimensions"] # vector dimensionalityprint(f"{dims}-dimensional vector")
Bucket Search
Search any Telnyx Storage bucket using natural language. Upload files, trigger server-side embedding, then run similarity search — the query embedding happens server-side too.
Search
# Search with default bucket (from config.json)
./search.py "what are the project requirements?"# Search a specific bucket
./search.py "meeting notes" --bucket my-bucket
# Get more results
./search.py "API rate limits" --num 10
# JSON output (for scripting)
./search.py "deployment steps" --json
# Custom timeout
./search.py "long query" --timeout 45
# Full content (no truncation)
./search.py "details" --full
Output Format
Results are ranked by certainty score with confidence indicators:
--- Result 1 [HIGH] (certainty: 0.923) ---
Source: docs/requirements.md
The project requires Python 3.8+ and a valid Telnyx API key...
--- Result 2 [MED] (certainty: 0.871) ---
Source: notes/planning.md
We discussed the requirements in the planning meeting...
from search import search, similarity_search
# Quick search (returns formatted text)print(search("your query", bucket_name="my-bucket"))
# Get structured results
results = similarity_search("your query", num_docs=5, bucket_name="my-bucket")
for doc in results.get("data", []):
print(doc["source"], doc["certainty"])
print(doc["content"][:200])
Index Content
Upload files to a Telnyx Storage bucket and trigger embedding so they become searchable.
Upload Files
# Upload a single file
./index.py upload path/to/file.md
# Upload to a specific bucket
./index.py upload path/to/file.md --bucket my-bucket
# Upload with a custom key (filename in bucket)
./index.py upload path/to/file.md --key docs/custom-name.md
# Upload all markdown files from a directory
./index.py upload path/to/dir/ --pattern "*.md"# Upload all files from a directory
./index.py upload path/to/dir/
Trigger Embedding
After uploading files, trigger the embedding process to make them searchable:
# Embed files in default bucket
./index.py embed
# Embed files in a specific bucket
./index.py embed --bucket my-bucket
Check Embedding Status
./index.py status <task_id>
List Files and Buckets
# List files in default bucket
./index.py list
# List files in a specific bucket
./index.py list --bucket my-bucket
# List files with a prefix filter
./index.py list --prefix docs/
# Show embedding status for a bucket
./index.py list --embeddings
# List all embedded buckets
./index.py buckets
Create a Bucket
./index.py create-bucket my-new-bucket
# With a specific region
./index.py create-bucket my-new-bucket --region us-central-1
