DOC / 01 Hosted provider
OpenAI.Hosted models, fixed destination.
Connect an OpenAI account while Palshell keeps the service endpoint fixed and the API key inside the operating system credential store.
01 Setup
Connect OpenAI without guesswork.
Make these changes in the installed Palshell desktop application. This website never collects provider credentials.
- 01Select OpenAI
Open Palshell, then go to Settings → AI Providers and select OpenAI.
- 02Enter the API key
Paste the key into the API key field. The Base URL is fixed by Palshell and cannot be edited.
- 03Choose a model
Keep Palshell’s current default
gpt-5.4, select Discover models, or enter another exact model ID available to your account. - 04Save and verify
Select Save, then Test connection.
- 05Make it the default if needed
Select Set default when OpenAI should be the first provider Palshell uses.
02 Exact configuration
Use these values. Keep credentials out of URLs.
https://api.openai.com/v1POST /v1/responsesModel discoveryGET /v1/models03 API cost
GPT-5.4 is billed by tokens used.
OpenAI bills the API account directly. Palshell does not add a provider surcharge or include API usage in a Palshell license.
- Input
- $2.50 / 1M tokens
- Cached input
- $0.25 / 1M tokens
- Output
- $15.00 / 1M tokens
Instructions and context sent to the model.
Eligible prompt content reused from cache.
Reasoning and response tokens generated by the model.
(input ÷ 1,000,000 × $2.50) + (output ÷ 1,000,000 × $15.00)Example: 10,000 input tokens and 2,000 output tokens cost about $0.055 before cached-input discounts or other feature charges.
Rates shown in USD and checked September 7, 2026. A GPT-5.4 request above 272,000 input tokens uses premium long-context rates, and optional provider tools can add separate charges. Other OpenAI models have different prices.
Current GPT-5.4 pricingThe destination stays visible. The secret stays native.
The API key is written directly to macOS Keychain or Windows Credential Manager. It is never stored in website or browser storage, Palshell’s application JSON, documentation, source code, analytics, or logs. Palshell retrieves it only for the immediate provider operation and never displays it again after saving.
04 Troubleshooting
If OpenAI does not connect, start here.
- 01API key not found
- Enter the key again in the OpenAI provider card, then select Save. Palshell checks only whether the credential item exists; it never reads the stored value back into the screen.
- 02Invalid or unavailable model ID
- Run Discover models and choose an ID returned for the authenticated account, or enter the exact case-sensitive model ID supplied by OpenAI.
- 03Test connection fails
- Confirm the credential and model, then retry. Also check network access, the provider service status, and whether the account may use that model.
05 Official references