Skip to main content
On an Enterprise account, Reagent sends all its AI requests to an endpoint your organization chooses instead of Reagent’s hosted models. The endpoint can be a model gateway (a service that passes requests on to AI models), a cloud provider or a model server on your own computer. If you’re the IT admin choosing the endpoint and model, start with For IT Admins.

Connect Your Endpoint

You need the Endpoint URL, API key and Model ID from your IT team, plus any options they want changed. If the endpoint is only reachable on your organization’s network, connect to that network or your VPN first. The Custom Endpoint window opens after you sign in (and, the first time, answer the welcome questions) whenever this computer has no saved endpoint: the first time you sign in on a computer, after you log out or your sign-in session ends, and after Clear. You can’t chat or close the window until an endpoint passes the test. To leave, click Sign out.
1

Enter the endpoint URL, API key and model ID

Paste the base URL, such as https://openrouter.ai/api/v1, without /chat/completions or /responses. Type the model ID exactly as your endpoint lists it, vendor prefix included (openai/gpt-5.4). Leave the other options on their defaults unless your IT team says otherwise.
2

Click Apply

Reagent tests the endpoint before it saves anything. On a pass, it shows Connected, any warnings and the API (request format) it will use, such as “Reagent will use your endpoint’s Chat Completions.” If the test fails, Reagent saves nothing and shows the error under Apply (see Troubleshooting).
3

Click Continue

You can start using Reagent.

Change or Clear Your Endpoint

  1. Open Settings > Endpoint.
  2. Under Connection, change what you need. The saved API key isn’t shown again. Fill in its field (“Saved API key (leave blank to keep)”) only to replace the key.
  3. Click Apply. Reagent re-tests the endpoint and, on a pass, shows the API it uses, such as “Connected — using Chat Completions.” Changes apply from the next request.
Once an endpoint is saved, the API in use line under Connection shows the API Reagent uses for your model, such as “Responses API — reasoning on” or “Chat Completions — reasoning off”. Notes under it explain why, for example that the endpoint doesn’t report token usage while streaming, or why reasoning is off and which API setting could turn it back on. The Supported inputs line shows what your model can be sent: “Text, image” when it takes pictures, and “Text” when it doesn’t. With a text-only model, Reagent turns pictures away when you attach them. If your endpoint rejects an option Reagent sent, or needs a different API for your model, Reagent retries once with adjusted settings and remembers the change. The API in use line then notes “Changed after the last test”. Click Apply to re-test. To remove the endpoint, go to Settings > Endpoint, click Clear next to Saved endpoint credentials, then click Clear credentials. Reagent forgets the URL, API key, model ID and options on this computer, and the Custom Endpoint window opens again. Your endpoint settings, API key included, are saved on this computer only, in Reagent’s app data folder (enterprise-config.enc). They’re encrypted with your operating system’s secure storage and aren’t synced. Reagent deletes them when you log out, when your sign-in session ends and when you click Clear. If the operating system can’t provide encryption, the Custom Endpoint window and Settings > Endpoint warn: “Encryption is unavailable on this system. Credentials will be stored as plain text in <userData>/enterprise-config.json.”

Fields and Options

Output limits and context sizes are measured in tokens, the small chunks of text a model reads and writes.
Reagent reads the model family from the part of the ID after the last / (openai/gpt-5.4 reads as gpt-5.4). The family sets which API Auto picks, which Reasoning effort values Reagent sends and the default Context window.Reagent can’t recognize a gateway alias that doesn’t name the model. For one, set API and Context window yourself. It also can’t recognize Claude IDs in other formats, such as us.anthropic.claude-…. For those, set Context window and Max output tokens yourself, and leave Reasoning effort on Model default.
Provider only changes how Reagent formats its tool definitions for Gemini models. Auto-detect reads the model ID: an ID containing claude counts as Anthropic, gpt as OpenAI, gemini as Gemini and anything else as Other. Choose Gemini only for a Gemini model whose ID doesn’t contain “gemini”.
Auto picks the API for each model after testing the endpoint. Chat Completions or Responses API always uses that one, and the test fails if you choose Responses API and the endpoint doesn’t provide it for your model.On Auto, recognized OpenAI reasoning models use the Responses API when your endpoint provides it for that model, and Chat Completions otherwise. On the Responses API, they can use tools with reasoning on and carry their reasoning between tool calls. Every other model uses Chat Completions. Recognized IDs start with gpt-5 (except -chat variants), gpt-6 or o followed by a digit (such as o3), contain codex, or are GPT models with -pro in the ID.
How much the model reasons before it answers. Reagent sends your choice only to models that accept that value, and others use their default. Two exceptions: IDs that start with claude (after any prefix) never get a reasoning effort on Chat Completions, and IDs Reagent doesn’t recognize get your choice unchanged.Your choice applies to the main chat and the Wwise subagent. For chat titles, long-chat summaries and the Analyst, OpenAI reasoning models and Gemini models get the lowest effort they accept, and other models get none.
Caps the length of each response. Empty uses the model’s default, except for IDs that start with claude on Chat Completions. For those, Reagent asks for up to 32,000 tokens, because some gateways otherwise apply a much lower limit. If a model rejects that, Reagent stops sending it.

Context Window

The context window is how much conversation the model can take in. Reagent uses it for the context ring next to the send button and to decide when to summarize a long chat. It doesn’t limit what your endpoint accepts. Empty uses a default based on the model ID: Set it when your model’s real limit differs, especially for local models with small windows. Too large a value lets a chat run past what the model accepts before Reagent summarizes it. Reagent starts summarizing at 50,000 tokens at the earliest, so choose a model whose window is well above that. If your endpoint doesn’t report token usage while streaming, the context ring’s tooltip adds “(estimated — this endpoint reports no usage)”.

For IT Admins

  • One model for everything. The model in Model ID handles every language-model request: chats, the Wwise subagent, the Analyst, chat titles and long-chat summaries.
  • Prompt caching. Each chat request repeats the previous one and adds to the end, so a provider that caches prompts reads most of every request from its cache. Main chat requests (not subagent, title or summary requests) carry one caching signal, chosen by model: If your endpoint rejects either field, Reagent sends the request again without it and leaves it out until the app restarts. If your gateway spreads requests across several provider accounts or regions, turn on sticky routing (session affinity) so a chat keeps reaching the same cache.
  • Responses API requests. Reagent sends the conversation with each request and sets store: false, so your endpoint isn’t asked to keep conversation state.

What the Endpoint Needs

Any OpenAI-compatible endpoint that meets these requirements works. The app’s compatibility note names TrueFoundry, LiteLLM, OpenRouter, Azure OpenAI, Together.ai, Ollama and vLLM. On Chat Completions, gpt-5.4 and later GPT-5 versions, and gpt-6, refuse tools unless reasoning is off, so Reagent turns reasoning off for them and shows a warning. gpt-6 can’t turn reasoning off, so it works only through the Responses API. For these models, make sure your gateway provides /responses.

What Apply Checks

Reagent checks the fields, then sends a few short test requests that don’t touch REAPER or your project:
  • A one-word reply on Chat Completions, to confirm the URL, key and model ID work. It’s skipped when API is Responses API.
  • Whether the endpoint provides the Responses API for this model, for recognized OpenAI reasoning models on Auto or when you chose Responses API.
  • A streamed request with one test tool, to see whether the model calls it.
  • A follow-up that checks reasoning carried between steps on the Responses API, or streamed token usage on Chat Completions.
  • A caching check: the same request, with a long fixed instruction, sent twice. If the second reply reports no cached tokens, Apply still succeeds and warns that every request will be billed at the full input price. A Gemini model gets no warning, because its caching isn’t guaranteed.
If both the Chat Completions and Responses paths answer “not found”, Reagent tries once more with /v1 added to your URL and keeps it if that works. Each step is retried once after a temporary failure, such as a timeout, a rate limit (429) or a server error (5xx). If a step still fails, Reagent saves nothing and shows your endpoint’s own error under Apply.

Troubleshooting

Some messages call the Apply button “Test & Apply”.
During setup, the error under Apply adds “This endpoint may only be reachable from your organization’s network or VPN.” In chat, an error box reads “Can’t reach your organization’s model endpoint.”Connect to your organization’s network or VPN, and check the Endpoint URL with your IT team. For a local server, check that it’s running at that address. Then click Apply again.If the test times out, the endpoint answered too slowly. A local server may still be loading the model, so wait and try again.
The URL doesn’t point at the endpoint’s OpenAI-compatible API. Use the base URL, usually ending in a version path such as /v1, without /chat/completions or /responses.
  • “Check the base URL and the model ID”: the endpoint may not serve that model ID.
  • “The Responses API is not available for this model on this endpoint”: set API to Auto or Chat Completions, or ask your gateway admin to enable /responses for the model.
  • “Check the endpoint URL and provider API path”: the URL returned a web page instead of an API response.
The message under Apply, or in a chat error box reading “Your organization’s model endpoint returned”, shows the status code, your endpoint’s own explanation and, when Reagent can explain the failure, a hint. Give the text under Endpoint response to your gateway admin.
The endpoint works and Reagent saved it, with a caveat:
  • “Your endpoint’s Responses API could not be used for this model (…), so Reagent uses Chat Completions.” You can leave it, or ask your gateway admin to enable /responses for the model and click Apply again.
  • “The model accepted tools but did not call one, so Reagent’s tools may not work with it.” Choose a model with stronger tool calling.
  • “This model only runs tools on Chat Completions with reasoning turned off, so Reagent will use it without reasoning.” To keep reasoning on, set API to Responses API if your endpoint provides it for this model, and click Apply again.
  • “Your endpoint’s Responses API returns reasoning without its encrypted content…” or “Your endpoint could not use replayed reasoning…” Reagent can’t carry the model’s reasoning between tool calls, so the model reasons afresh at each step. There’s nothing to do.
The message reads “The endpoint passed the test, but Reagent could not save its settings” and gives the reason. If another program had the file open, click Apply again. If it keeps failing, send the full message to Reagent support.
If the error says the model refuses Reagent’s tools on Chat Completions while reasoning is on, or “can only run Reagent’s tools through the Responses API”, ask your gateway admin to enable /responses for the model. Then set API to Auto (or Responses API if Reagent doesn’t recognize your model ID) and click Apply.
The response probably hit an output limit. If you set Max output tokens, raise or clear it. If it’s empty, your gateway may apply its own low default, so set a value your model supports. If the model rejects your value (“The model rejected the output-token cap. Lower or clear Max output tokens.”, or a message naming the value Reagent sent from your settings), lower or clear it.
Reagent may be assuming a larger window than your model has. Set Context window to the model’s real limit and click Apply. Or use a model with a larger window, or start a new chat (see Context Window).

Enterprise Overview

How Enterprise accounts work and which features differ.

Privacy & Security

Where your conversation goes on an Enterprise account.