ktxd 0.1.1

A local Responses API compatibility proxy for Codex CLI and Chat Completions backends.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

[server]
bind = "127.0.0.1:3000"

[models.DeepSeek-V4-Pro]
# Public model slug accepted from Codex requests and exposed by /v1/models.
public_model = "DeepSeek-V4-Pro"
upstream_family = "chat_completions"

# Azure AI Foundry deployment/model binding. Keep these separate from the public
# slug because many Azure endpoints bind deployment in the URL while others also
# accept or require a model/deployment identifier in the request JSON body.
upstream_deployment = "DeepSeek-V4-Pro"
upstream_model = "DeepSeek-V4-Pro"
chat_completions_url = "https://YOUR-RESOURCE.YOUR-REGION.services.ai.azure.com/openai/v1/chat/completions"

# Secret is read from the environment at request time and is never logged.
auth_header = "api-key"
auth_env_var = "AZURE_AI_FOUNDRY_API_KEY"

# Azure compatibility toggles.
send_model_in_body = true
include_stream_usage = true
retry_without_stream_options_on_4xx = true
instruction_role = "system"

display_name = "DeepSeek V4 Pro"
description = "DeepSeek-V4-Pro via Azure AI Foundry Chat Completions"
context_window = 1000000