# Endpoints that keep no data

`GET /v1/endpoints/zdr` lists every endpoint that keeps no data from your requests: one entry per
model on sale and provider serving it.

## Request

Send it with an API key. It takes no field. The [models guide](/docs/models) says how a model is
named and served; [One model's providers](/docs/reference/endpoints) answers every endpoint of
one model, whether it keeps data or not.

Souk Bot answers the customers of a clothing shop in Oran, and Yacine wants none of their messages
kept by a provider. He reads this list and keeps Souk Bot on the models it names.

```python
import json
import os
import urllib.request

API_KEY = os.environ["OPENDUNES_API_KEY"]
ADDRESS = "https://opendunes.com/v1/endpoints/zdr"

request = urllib.request.Request(
    ADDRESS,
    headers={"Authorization": "Bearer " + API_KEY},
)
with urllib.request.urlopen(request) as response:
    endpoints = json.load(response)["data"]

for endpoint in endpoints:
    print(endpoint["model_id"], endpoint["provider_name"])
```

```javascript
const API_KEY = process.env.OPENDUNES_API_KEY;
const ADDRESS = "https://opendunes.com/v1/endpoints/zdr";

const response = await fetch(ADDRESS, {
  headers: { Authorization: `Bearer ${API_KEY}` },
});
const { data } = await response.json();

for (const endpoint of data) {
  console.log(endpoint.model_id, endpoint.provider_name);
}
```

```go
package main

import (
	"encoding/json"
	"fmt"
	"net/http"
	"os"
)

const address = "https://opendunes.com/v1/endpoints/zdr"

func main() {
	apiKey := os.Getenv("OPENDUNES_API_KEY")

	req, err := http.NewRequest("GET", address, nil)
	if err != nil {
		panic(err)
	}
	req.Header.Set("Authorization", "Bearer "+apiKey)

	res, err := http.DefaultClient.Do(req)
	if err != nil {
		panic(err)
	}
	defer res.Body.Close()

	var body struct {
		Data []struct {
			ModelID      string `json:"model_id"`
			ProviderName string `json:"provider_name"`
		} `json:"data"`
	}
	if err := json.NewDecoder(res.Body).Decode(&body); err != nil {
		panic(err)
	}
	for _, endpoint := range body.Data {
		fmt.Println(endpoint.ModelID, endpoint.ProviderName)
	}
}
```

```bash cURL
ADDRESS="https://opendunes.com/v1/endpoints/zdr"

curl "$ADDRESS" \
  -H "Authorization: Bearer $OPENDUNES_API_KEY"
```

## Response

`200 OK`.

| Field | Holds |
|---|---|
| `data` | every endpoint that keeps no data, one per model on sale and provider |
| `name` | in each: the provider and the model it serves there |
| `model_id` | in each: the model's ID, `lab/model` |
| `model_name` | in each: the model's name on this provider's list |
| `provider_name` | in each: the provider |
| `tag` | in each: the provider's tag: send it in `provider.only` or `provider.order` to pick this endpoint |
| `context_length` | in each: the most tokens a request and its answer hold together here |
| `max_completion_tokens` · `max_prompt_tokens` | in each: the longest answer and the longest prompt, in tokens, or `null` |
| `pricing` | in each: this provider's prices in DA, one per meter, each a string per unit of that meter. Beside them, `discount`: the promotion already inside those prices, `0.5` for half off, `0` when there is none |
| `quantization` | in each: the precision the model runs at here, such as `fp8`, or `unknown` |
| `supported_parameters` | in each: the request fields this provider honours |
| `supports_tool_choice` | in each: which `tool_choice` values it honours: `none`, `auto`, `required`, `function`, each `true` or `false` |
| `supports_implicit_caching` | in each: `true` when it caches a repeated prompt without being asked |
| `native_tools` | in each: the helpers this provider runs with its own built-in tool, keyed by the helper's name, `opendunes:‹name›`; empty when it has none |
| `supports_voice_cloning` · `supports_multiple_audio_references` · `supports_image_reference` | in each: on a speech model, whether it takes a reference voice clip, several clips, or a picture describing the voice |
| `status` | in each: `0` when it is serving normally; below `0`, degraded or down |
| `uptime_last_5m` · `uptime_last_30m` · `uptime_last_1d` | in each: the share of its requests that succeeded, in percent, or `null` |
| `latency_last_30m` · `throughput_last_30m` | in each: time to the first token in milliseconds, and tokens per second, at `p50`, `p75`, `p90` and `p99`, or `null` |

## What it cost

Free. Reading the catalogue is never charged.

## Errors

`error.code` carries the status, and `error.message` says why.

| Status | When |
|---|---|
| 401 | the key is missing, wrong, disabled or expired |
| 403 | the key answers only from the addresses set on it, and this request came from another |
| 503 | the list could not be read just now; `Retry-After` says how many seconds to wait |
