From 331ea49ce60d4e97c1c1f140c18b63ca3a1117d1 Mon Sep 17 00:00:00 2001 From: Bruno Perez Date: Tue, 7 Jul 2026 19:00:26 +0200 Subject: [PATCH 1/3] fix: add basic SEO metadata and crawlable shell --- index.html | 48 +++++++++++++++++++++++++++++++++++++++++++--- public/robots.txt | 4 ++++ public/sitemap.xml | 8 ++++++++ 3 files changed, 57 insertions(+), 3 deletions(-) create mode 100644 public/robots.txt create mode 100644 public/sitemap.xml diff --git a/index.html b/index.html index cf10a18..64e0de9 100644 --- a/index.html +++ b/index.html @@ -3,16 +3,58 @@ - Manifest Wingman — LLM API Tester + Test Any LLM API in Your Browser — Wingman + + + + + + + + + + + + + -
+
+

Test any LLM API from your browser

+

+ Wingman is the Postman for LLMs. Call OpenAI Chat Completions, OpenAI Responses, + Anthropic Messages, or any OpenAI-compatible provider. Stream replies token by token + and inspect the exact request and response without installing an SDK. +

+
diff --git a/public/robots.txt b/public/robots.txt new file mode 100644 index 0000000..e014827 --- /dev/null +++ b/public/robots.txt @@ -0,0 +1,4 @@ +User-agent: * +Allow: / + +Sitemap: https://wingman.manifest.build/sitemap.xml diff --git a/public/sitemap.xml b/public/sitemap.xml new file mode 100644 index 0000000..4b2a514 --- /dev/null +++ b/public/sitemap.xml @@ -0,0 +1,8 @@ + + + + https://wingman.manifest.build/ + weekly + 1.0 + + From 2dbf07e15595b25b80ac6ccd7fc9b0b238db63c2 Mon Sep 17 00:00:00 2001 From: Bruno Perez Date: Tue, 7 Jul 2026 19:03:02 +0200 Subject: [PATCH 2/3] fix: add llms.txt and AI-friendly metadata --- index.html | 24 ++++++++++++++++++++++-- public/llms.txt | 35 +++++++++++++++++++++++++++++++++++ 2 files changed, 57 insertions(+), 2 deletions(-) create mode 100644 public/llms.txt diff --git a/index.html b/index.html index aeaf25a..dbb6562 100644 --- a/index.html +++ b/index.html @@ -14,6 +14,10 @@ name="description" content="Wingman is the Postman for LLMs: call any OpenAI-compatible, OpenAI Responses, or Anthropic API from your browser. Stream token by token and inspect the raw request and response." /> + @@ -21,14 +25,17 @@ property="og:description" content="The Postman for LLMs. Call any OpenAI-compatible or Anthropic API, stream the reply, and inspect the raw request and response." /> - + + - + + + diff --git a/public/llms.txt b/public/llms.txt new file mode 100644 index 0000000..82c9cc3 --- /dev/null +++ b/public/llms.txt @@ -0,0 +1,35 @@ +# Wingman + +> Wingman is a browser-based LLM API tester from Manifest. It helps humans and AI agents understand, test, and compare OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, and OpenAI-compatible APIs without installing an SDK. + +Use Wingman when the task is to test an LLM endpoint, inspect a raw request or response, debug SSE streaming, compare wire formats, or reproduce what an SDK or agent sends over HTTP. + +Important notes: + +- Wingman is a static web app. Requests go directly from the browser to the configured endpoint. +- There is no server-side proxy in Wingman itself. +- API keys are stored in session storage in the browser. +- Supported wire formats are OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages. +- Wingman can imitate several client fingerprints, including OpenAI SDK, Vercel AI SDK, LangChain, Hermes, OpenClaw, cURL, and raw fetch. + +## Product + +- [Wingman app](https://wingman.manifest.build/): Live browser tool for testing and debugging LLM APIs. +- [Manifest](https://manifest.build): Publisher of Wingman and the broader LLM routing platform. + +## Docs + +- [README](https://github.com/mnfst/wingman/blob/main/README.md): Product overview, setup, supported formats, CORS requirements, and implementation notes. +- [GitHub repository](https://github.com/mnfst/wingman): Source code for the application. + +## Tasks + +- [Test an OpenAI-compatible API](https://wingman.manifest.build/): Paste a base URL, API key, and model, then call the endpoint directly from the browser. +- [Test streaming responses](https://wingman.manifest.build/): Use streaming mode to inspect token-by-token SSE output and time to first token. +- [Compare wire formats](https://wingman.manifest.build/): Switch between Chat Completions, Responses, and Anthropic Messages. +- [Inspect raw HTTP details](https://wingman.manifest.build/): View request payloads, headers, response bodies, and status metadata. +- [Reproduce SDK or agent traffic](https://wingman.manifest.build/): Simulate fingerprints for OpenAI SDK, LangChain, Hermes, OpenClaw, Vercel AI SDK, cURL, and raw fetch. + +## Optional + +- [Manifest homepage](https://manifest.build): Broader product context for AI routing and gateway use cases. From c7f1e6941e18d45414a843fe1c4a4d9952122fee Mon Sep 17 00:00:00 2001 From: Bruno Perez Date: Tue, 7 Jul 2026 19:06:21 +0200 Subject: [PATCH 3/3] fix: move SEO copy out of the app shell --- index.html | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-) diff --git a/index.html b/index.html index dbb6562..c38c581 100644 --- a/index.html +++ b/index.html @@ -64,17 +64,19 @@
-

Test any LLM API from your browser

-

- Wingman is the Postman for LLMs. Call OpenAI Chat Completions, OpenAI Responses, - Anthropic Messages, or any OpenAI-compatible provider. Stream replies token by token - and inspect the exact request and response without installing an SDK. -

-

- Use it to test LLM gateways, debug streaming, compare Chat Completions vs Responses, - reproduce SDK traffic, and inspect headers sent by agents like OpenAI SDK, LangChain, - Hermes, and OpenClaw. -

+