Skip to content

fix(ai): tolerate a full chat endpoint as a custom provider base URL when refreshing models #578

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
hoishing wants to merge 1 commit into
base: main
Choose a base branch
from
+11 −2
Open

fix(ai): tolerate a full chat endpoint as a custom provider base URL when refreshing models #578

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 11 additions & 2 deletions Sources/Fluid/Services/ModelRepository.swift
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -249,8 +249,17 @@ final class ModelRepository {

let isAnthropic = providerID == "anthropic" || baseURL.contains("anthropic.com")

// Construct the models endpoint URL
let urlString = baseURL.hasSuffix("/") ? "\(baseURL)models" : "\(baseURL)/models"
// Construct the models endpoint URL. Users frequently paste the full chat
// endpoint (…/chat/completions or …/responses) as the base URL; append `/models`
// to that verbatim and it 404s. Strip a trailing endpoint segment (and slash) so
// the list resolves to <root>/models.
var normalizedBase = baseURL.trimmingCharacters(in: .whitespacesAndNewlines)
for suffix in ["/chat/completions", "/responses"] where normalizedBase.hasSuffix(suffix) {
normalizedBase = String(normalizedBase.dropLast(suffix.count))
break
}
while normalizedBase.hasSuffix("/") { normalizedBase = String(normalizedBase.dropLast()) }
Comment on lines +257 to +261

@chatgpt-codex-connector chatgpt-codex-connector Bot Jul 11, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Normalize trailing slashes before matching endpoint suffixes

If a saved custom provider URL is copied with a trailing slash, e.g. https://host/v1/chat/completions/ or https://host/v1/responses/, this suffix check runs before the trailing slash is removed, so neither endpoint suffix matches. The code then appends /models to .../chat/completions or .../responses, preserving the same 404 path the change is meant to fix; trim trailing slashes before the suffix loop or repeat the suffix check afterward.

Useful? React with 👍 / 👎.

let urlString = "\(normalizedBase)/models"
Comment on lines +256 to +262

@greptile-apps greptile-apps Bot Jul 11, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Trailing slash defeats endpoint stripping

Trailing slash stripping happens after the suffix check, so a URL ending in /chat/completions/ (with trailing slash) won't match hasSuffix("/chat/completions") and gets appended with /models unchanged — producing …/chat/completions/models and the same 404 as before. Reversing the two steps (strip slashes first, then check for endpoint suffixes) fixes this. Compare with LLMClient.endpoint, which uses .contains instead of .hasSuffix and is immune to a trailing slash.

Suggested change
var normalizedBase = baseURL.trimmingCharacters(in: .whitespacesAndNewlines)
for suffix in ["/chat/completions", "/responses"] where normalizedBase.hasSuffix(suffix) {
normalizedBase = String(normalizedBase.dropLast(suffix.count))
break
}
while normalizedBase.hasSuffix("/") { normalizedBase = String(normalizedBase.dropLast()) }
let urlString = "\(normalizedBase)/models"
var normalizedBase = baseURL.trimmingCharacters(in: .whitespacesAndNewlines)
while normalizedBase.hasSuffix("/") { normalizedBase = String(normalizedBase.dropLast()) }
for suffix in ["/chat/completions", "/responses"] where normalizedBase.hasSuffix(suffix) {
normalizedBase = String(normalizedBase.dropLast(suffix.count))
break
}
while normalizedBase.hasSuffix("/") { normalizedBase = String(normalizedBase.dropLast()) }
let urlString = "\(normalizedBase)/models"
Prompt To Fix With AI 
This is a comment left during a code review.
Path: Sources/Fluid/Services/ModelRepository.swift
Line: 256-262

Comment:
**Trailing slash defeats endpoint stripping**

Trailing slash stripping happens _after_ the suffix check, so a URL ending in `/chat/completions/` (with trailing slash) won't match `hasSuffix("/chat/completions")` and gets appended with `/models` unchanged — producing `…/chat/completions/models` and the same 404 as before. Reversing the two steps (strip slashes first, then check for endpoint suffixes) fixes this. Compare with `LLMClient.endpoint`, which uses `.contains` instead of `.hasSuffix` and is immune to a trailing slash.

```suggestion
        var normalizedBase = baseURL.trimmingCharacters(in: .whitespacesAndNewlines)
        while normalizedBase.hasSuffix("/") { normalizedBase = String(normalizedBase.dropLast()) }
        for suffix in ["/chat/completions", "/responses"] where normalizedBase.hasSuffix(suffix) {
            normalizedBase = String(normalizedBase.dropLast(suffix.count))
            break
        }
        while normalizedBase.hasSuffix("/") { normalizedBase = String(normalizedBase.dropLast()) }
        let urlString = "\(normalizedBase)/models"
```

How can I resolve this? If you propose a fix, please make it concise.

Fix in Codex

guard let url = URL(string: urlString) else {
DebugLogger.shared.error(
"fetchModels: Invalid URL constructed from baseURL='\(baseURL)' -> '\(urlString)'",
Expand Down
Loading