---
title: "KV: Global Key-Value Storage for Edge Workloads | Telnyx"
canonical: https://telnyx.com/products/kv
---
# KV: Global Key-Value Storage for Edge Workloads | Telnyx

**TELNYX EDGE COMPUTE**

# KV: Key-value storage at edge speed

Globally distributed storage built for read-heavy edge workloads. No database to manage, no infrastructure to provision. Just write, read, and scale.

[Start Building](https://telnyx.com/sign-up)[CONTACT US](https://telnyx.com/contact-us)

WHY TELNYX

## Not just another key-value store

Most key-value stores lock you into a single compute platform or bury you in database-style complexity. Telnyx KV does neither. It works from anywhere, runs colocated with Telnyx telecom infrastructure, and ships as part of a broader edge compute platform. One account, one API, one bill.

* ### One platform for telecom and compute  
Voice, messaging, storage, compute, and inference on a single API. Your application talks to the same infrastructure that carries the call, stores the data, and runs the model. No patchwork of vendors to reconcile.
* ### Private network, not the public internet  
Traffic between your functions, storage, and telephony stays on the Telnyx private network. Lower latency, no exposure to public internet routing, no hop through a third-party cloud.
* ### Carrier-grade reliability  
Telnyx infrastructure carries production voice and messaging traffic for enterprises globally. The same network that handles millions of calls powers your edge compute workloads. Tier-1 backbone, redundant routing, 24/7 NOC.

FEATURES

## Production-ready storage for the patterns you already use

Two access paths, one store. Zero-credential access from edge functions, REST for everything else.

* Zero-credential edge access  
Your code reaches KV directly from edge functions. No API keys to manage, no extra network hop. The store is available the moment your function runs.
* Works from any language  
Read and write from Python, Go, Rust, curl, or anything that speaks HTTP. The REST API is a core access path, not an afterthought.
* Store any format  
JSON, protobuf, raw text, binary. No schema to define, no envelope around your data, no opinions about your format. You choose the serialization.
* Automatic expiration  
Set a time-to-live on any key and the platform handles the rest. No cleanup jobs, no stale data, no cron-based garbage collection.
* Global by default  
Write from Chicago, read from Singapore. KV is accessible from any edge function in any region, not pinned to a single zone.
* One store, every path  
Edge functions and backends share the same data. Write from your function, read from your server. No replication step, no sync lag.

HOW IT WORKS

## Three steps to your first value

No schemas, no provisioning, no configuration. Create a namespace, write a key, read it back.

[Read the docs](https://developers.telnyx.com/docs/edge-compute/kv)

cURL

```
# Write a value
curl -X PUT "https://api.telnyx.com/v2/storage/kvs/{namespace_id}/keys/session_123" \
  -H "Authorization: Bearer $TELNYX_API_KEY" \
  -H "Content-Type: application/octet-stream" \
  -d '{"user_id": "42", "role": "admin"}'

# Read it back
curl "https://api.telnyx.com/v2/storage/kvs/{namespace_id}/keys/session_123" \
  -H "Authorization: Bearer $TELNYX_API_KEY"
```

BINDINGS

## What connects to KV.

[See all products](https://telnyx.com/products)

* ### Functions  
Edge functions read and write KV through a typed env binding. No API key, no network hop. The most natural access path for serverless code at the edge.  
[Learn more](https://telnyx.com/products/functions)
* ### Stateful Actors  
Actors use KV for shared state across instances. Session data written by a function is readable from an actor, and vice versa. One store, no sync.  
[Learn more](https://telnyx.com/products/stateful-actors)
* ### Object Storage  
Store and retrieve objects from the same edge locations as your functions. Co-located storage means lower latency for your AI workloads.  
[Learn more](https://telnyx.com/products/cloud-storage)
* ### Inference  
Run LLM inference on GPUs in the same facility as your KV store. No network hops between your data layer and your model serving.  
[Learn more](https://telnyx.com/products/inference)

## Sign up and start building.

Get an API key, create a namespace, and write your first value in under five minutes.

[SIGN UP](https://telnyx.com/sign-up)[TALK TO US](https://telnyx.com/contact-us)

## Sign up and start building.

Get an API key, create a namespace, and write your first value in under five minutes.

[SIGN UP](https://telnyx.com/sign-up)[TALK TO US](https://telnyx.com/contact-us)

## FAQ

### What is Telnyx KV?

Telnyx KV is a globally distributed key-value store built for read-heavy edge workloads. You store data under a key and read it back fast from anywhere. It's designed for session data, cached responses, feature flags, and small values your applications need on every request.

### How is KV different from a database?

### Can I use KV without edge functions?

### What data formats does KV support?

### How does automatic expiration work?

### Is data written from edge functions immediately readable via REST?

### What is the difference between KV and Object Storage?

### How do I get started?

[Start building](https://telnyx.com/sign-up)