Guide
Plans
Pool your team's ChatGPT and Claude plans beside your API keys. Every tool pointed at the gateway rotates across them, and you watch every plan's limits live.
pro_required means the organization is not on Pro yet.What a plan is
A plan is a ChatGPT (Plus, Pro, Team or Enterprise seat) or Claude (Pro, Max or Team seat) sign-in, added as one account in your organization's pool. Requests for that provider's models can then run on the plan's allowance instead of an API key. Each plan has a 5-hour and a weekly usage window; when one runs out, traffic moves to your next plan or key and the plan comes back on its own after the reset.
Add plans in the dashboard
- Open Settings › Connections › Plans and choose Add a plan.
- Choose whom it serves: Shared pool (every key in the organization) or Only my API keys (only keys you created). Any member can add a plan for their own keys; adding one to the shared pool needs an organization admin.
- ChatGPT plan: sign in through the browser (the tab ends on a
localhost:1455page that does not load; paste its address back) or import the sign-in Codex already holds. - Claude plan: approve on claude.ai. Depending on the deployment, the sign-in either comes straight back here or ends on a page with a one-time code you paste into the form; you can also import the sign-in Claude Code already holds on this machine. It appears only where the deployment has Claude plans enabled.
- Repeat for every account. Each plan shows as a card with its live limits.
Or let your coding agent do it
Paste this into your coding agent. It uses your organization key (EXPLABS_API_KEY), imports Codex sign-ins with your permission, and hands you the sign-in links it cannot open itself.
I pasted this into you myself, so add my ChatGPT and Claude plans to the Plans pool on myExperiential Labs gateway. Every tool I point at the gateway then rotates across these plans and my API keys, and I can watch the limits at https://pr-1347.preview.experientiallabs.ai/settings/connections#plans .This is my consent to wire it up. Plans are part of the Pro plan: a 402 with codepro_required means my org needs Pro at https://pr-1347.preview.experientiallabs.ai/credits first.Use my Experiential Labs org key as the bearer token on every call below. It is asecret: read it from the EXPLABS_API_KEY environment variable and never print it infull. If it is not set, ask me for it (I can mint one at https://pr-1347.preview.experientiallabs.ai/api-keys).First find my org id and the accounts it already has:GET https://api-pr-1347.preview.experientiallabs.ai/api/whoami with header Authorization: Bearer $EXPLABS_API_KEYCapture org_id, then GET https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections.Rows whose credential_kind is chatgpt_plan or claude_plan are plans. Pick a newsetup_alias for each plan you add (chatgpt-1, chatgpt-2, claude-1, ...), unusedacross ALL of my accounts.Add my ChatGPT plan (the Plus, Pro, Team or Enterprise seat I use Codex with):1. Ask me which way I prefer, then do exactly one:a. IMPORT (fastest, but it moves my sign-in): if this machine's Codex is signed in tothat account, read ~/.codex/auth.json in full (the whole file is the auth_json value):POST https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/openai/plans/importwith body {"auth_json": "<the file's contents as a string>", "setup_alias": "chatgpt-<n>"}Tell me clearly: my local Codex is now signed out and must run 'codex login' again.b. BROWSER (keeps both signed in): POST https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/openai/plans/sign-inwith body {"setup_alias": "chatgpt-<n>"} and give me the authorize_url from the response.I open it, approve on the ChatGPT account, and my browser then lands on alocalhost:1455 page that does not load. I paste you that page's full address, and youPOST https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/openai/plans/sign-in/completewith body {"state": "<state from the sign-in response>", "callback": "<the address I pasted>","setup_alias": "chatgpt-<n>"}Add my Claude plan (the Pro, Max or Team seat I use Claude with):0. First GET https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/anthropic/plans . If "available" is false, Claude plans are notenabled here: skip this and tell me so. Otherwise remember "completion" for below:"paste" means each sign-in ends on a page with a one-time code I paste back to you;"redirect" means it finishes on its own.1. Ask me which way I prefer, then do exactly one:a. IMPORT (fastest, but it moves my sign-in, and ONLY when completion is "paste"): ifthis machine's Claude Code is signed in to that account, read~/.claude/.credentials.json and POST its full contents tohttps://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/anthropic/plans/importbody {"auth_json": "<the file's contents as a string>", "setup_alias": "claude-<n>"}Tell me clearly: my local Claude Code is now signed out and must sign in again.On macOS the file may live in the Keychain instead; if it does, use the browser way.b. BROWSER (keeps both signed in): POST https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/anthropic/plans/sign-inbody {"setup_alias": "claude-<n>"} and give me the authorize_url from the response.I open it in a browser where I am signed in to claude.ai and approve.- If completion is "redirect", it finishes on its own and lands on my Plans page athttps://pr-1347.preview.experientiallabs.ai/settings/connections#plans .- If completion is "paste", the page ends with a one-time code. I paste it to you,and you POST https://api-pr-1347.preview.experientiallabs.ai/api/orgs/<org_id>/provider-connections/anthropic/plans/sign-in/completebody {"state": "<state from the sign-in response>", "callback": "<the code I pasted>","setup_alias": "claude-<n>"}After each add, report from the response: setup_alias, plan.label, plan.plan_type, and theused_percent of plan.primary and plan.secondary if present. A 409 means that sign-inexpired or was already used: start it again, do not retry the completion. Never print atoken, the credentials file contents, or my org key.Nothing else to configure: any harness (Claude Code, Codex, OpenCode, Cursor, ...) pointedat https://api-pr-1347.preview.experientiallabs.ai/v1 with my key uses the plans. Onelimitation to know: a request that sets an output ceiling (max_tokens) skips ChatGPT planrungs, so that case falls to my API keys.If you cannot do this, tell me and I will add the plan at https://pr-1347.preview.experientiallabs.ai/settings/connections#plans .
Use them from any harness
There is nothing plan-specific to configure. Point your tool at the gateway with your key as usual; requests for OpenAI and Anthropic models rotate across your plans and keys.
# Claude Code (Anthropic Messages)export ANTHROPIC_BASE_URL="https://api-pr-1347.preview.experientiallabs.ai"export ANTHROPIC_API_KEY="xpl_..."claude# Codex, OpenCode, Cursor, or any OpenAI-compatible toolexport OPENAI_BASE_URL="https://api-pr-1347.preview.experientiallabs.ai/v1"export OPENAI_API_KEY="xpl_..."
See Coding agents for per-tool setup.
Rotation and live limits
- Each plan card shows its 5-hour and weekly windows and when each resets, updating every minute while the page is open.
In rotationmeans the pool may route to the plan;Resting until resetmeans a window is used up and traffic goes elsewhere until it resets.Pausedmeans the plan is out of rotation until someone resumes it from its card; a paused plan never serves a request. An admin can pause any plan; a member can pause their own.- The pool's order is the same as for API keys. A plan that hits its limit hands the request to your next plan without an error.
API Fallback, top right of the Plans section, decides what happens once every plan is paused. On (the default), requests for a model your plans serve fall back to Experiential-funded provider routes, billed to credits. Off, they stay on your plans and answer 429 with the reset time. Only organization admins can change it.
Good to know
- Importing a Codex sign-in hands it over: run
codex loginagain for Codex itself. - A request that sets an output ceiling (
max_tokens) skips ChatGPT plans, which do not accept one. - A ChatGPT plan serves GPT-5 family models; a Claude plan serves Claude models.
- Tokens are stored encrypted and never returned; responses show a masked email only.
- Using a plan through a gateway is subject to the provider's terms for that plan.
API
Every call takes your organization key as a bearer token. Adding a plan needs Pro; reading and disconnecting never do.
| Endpoint | Purpose |
|---|---|
| GET /api/orgs/{org_id}/provider-connections/{openai|anthropic}/plans | Whether that provider's plans can be added on this deployment. |
| POST …/{openai|anthropic}/plans/sign-in | Start a browser sign-in: returns authorize_url and state. |
| POST …/{openai|anthropic}/plans/sign-in/complete | Finish it with the state and the address the browser landed on. |
| POST …/openai/plans/import | Import a Codex sign-in (the contents of ~/.codex/auth.json). |
| POST …/{openai|anthropic}/plans/{setup_alias}/refresh-usage | Re-read one plan's 5-hour and weekly windows now. |
| GET /api/orgs/{org_id}/provider-connections/accounts/usage | Every account, plans included, with pool state and usage. |
| DELETE /api/orgs/{org_id}/provider-connections/{provider}?setup_alias=… | Disconnect one plan; its stored sign-in is deleted. |