/models command returns only partial model list — missing pagination support

[nsfoxer]

Description

The /models command and background catalog refresh issue a single GET /v1/models request without handling OpenAI-style cursor pagination (has_more / after), so only the first page of results is returned. When a provider has more models than the default page size, most models are silently dropped.

Steps to reproduce

1. Configure a provider with a large model roster (e.g. OpenCode Go)
2. Run /models or codewhale models
3. Observe that the returned list is much shorter than the provider actually offers

Expected behavior

All available models from the provider are listed.

Actual behavior

Only the first page (~10–20 models) is returned; remaining models are missing.

Impact

Always happens for providers whose model count exceeds one page. Most models are unavailable for selection.

Environment

• OS: Windows 11
• codewhale version: 0.9.13
• Install method: source build
• Model/provider: OpenCode Go (and any provider with > default page size)
• Terminal app: Windows Terminal
• Shell: PowerShell

Logs, screenshots, or recordings

Relevant code in crates/tui/src/client.rs:

• ModelsListResponse struct (~line 3319) only has a data field — missing has_more and last_id
• list_models() (~line 2432) and fetch_catalog_delta() (~line 2514) both issue a single request with no pagination loop

OpenAI /v1/models paginated response shape:

{
  "object": "list",
  "data": [...],
  "has_more": true,
  "last_id": "model-xxx"
}

When has_more is true, the client must request the next page with ?after=<last_id> until has_more becomes false.

[github-actions]

Thanks @nsfoxer for the report.

This issue is staying open for maintainer triage. CodeWhale gets better because people bring us real edge cases from real machines, providers, regions, and workflows.

If you can add a reproduction, logs, version output, screenshots, or the provider/model involved, that makes it much easier for us to verify and harvest the fix. Maintainers may comment /lgtmi to mark recurring issue reporters as approved so this intake note is skipped next time.

[Hmbown]

Description

The /models command and background catalog refresh issue a single GET /v1/models request without handling OpenAI-style cursor pagination (has_more / after), so only the first page of results is returned. When a provider has more models than the default page size, most models are silently dropped.

Steps to reproduce

1. Configure a provider with a large model roster (e.g. OpenCode Go)
2. Run /models or codewhale models
3. Observe that the returned list is much shorter than the provider actually offers

Expected behavior

All available models from the provider are listed.

Actual behavior

Only the first page (~10–20 models) is returned; remaining models are missing.

Impact

Always happens for providers whose model count exceeds one page. Most models are unavailable for selection.

Environment

• OS: Windows 11
• codewhale version: 0.9.13
• Install method: source build
• Model/provider: OpenCode Go (and any provider with > default page size)
• Terminal app: Windows Terminal
• Shell: PowerShell

Logs, screenshots, or recordings

Relevant code in crates/tui/src/client.rs:

• ModelsListResponse struct (~line 3319) only has a data field — missing has_more and last_id
• list_models() (~line 2432) and fetch_catalog_delta() (~line 2514) both issue a single request with no pagination loop

OpenAI /v1/models paginated response shape:

{
  "object": "list",
  "data": [...],
  "has_more": true,
  "last_id": "model-xxx"
}

When has_more is true, the client must request the next page with ?after=<last_id> until has_more becomes false.

Thank you!! Fixing in next release