@callstack/apex 0.4.1 → 0.4.3

package/README.md

@@ -1,10 +1,25 @@
 # Apex CLI
 
 Configures the AI coding assistants on your machine to use
-[Callstack Apex](https://apex.callstack.com/) (`callstack/Apex`).
+[Callstack Apex](https://apex.callstack.com/) through Vercel AI Gateway (`callstack/apex`) or the
+Callstack portal (`callstack/Apex`).
 
-`apex init` detects the assistants on your machine, lets you pick which ones to configure,
-shows every change it wants to make, and only then writes. Nothing changes until you say so.
+`apex init` asks which gateway serves Apex, detects the assistants on your machine, lets you pick
+which ones to configure, shows every change it wants to make, and only then writes. Nothing
+changes until you say so.
+
+## Gateways
+
+| Gateway | `--provider` | Model | Key |
+| --- | --- | --- | --- |
+| Vercel AI Gateway | `vercel` | `callstack/apex` | `AI_GATEWAY_API_KEY` |
+| Callstack portal | `callstack` | `callstack/Apex` | `CALLSTACK_AUTH_TOKEN` |
+
+`apex init` asks which one to use. Without `--provider`, the default is the gateway whose key is
+already set in your shell, and Vercel AI Gateway when neither or both are. `apex detect` and
+`apex run` pick the gateway the same way, so `apex run codex` launches whatever you set up last as
+long as only that key is exported. Each gateway gets its own provider entry in every config, so
+switching gateways adds a provider next to the old one instead of rewriting it.
 
 ## Install
 
@@ -30,25 +45,26 @@ undo your configs, so run `apex undo` first if you want them back.
 | `apex init --apply` | Writes the planned changes without asking | Yes |
 | `apex undo` | Reverts the most recent Apex CLI setup, with the same preview and confirmation | Only after you confirm |
 | `apex undo --list` | Shows recorded setups and which are already undone (`--json` prints the journal) | No |
-| `apex run <assistant> [-- <args>]` | Launches the assistant with the gateway, model and credentials in the child environment | No |
+| `apex run [--provider <id>] <assistant> [-- <args>]` | Launches the assistant with the gateway, model and credentials in the child environment | No |
 | `apex completion <zsh\|bash\|fish>` | Prints a shell completion script | No |
 
 Each command accepts only its own flags, and `--help` and the shell completions list exactly those:
 
 | Command | Flags |
 | --- | --- |
-| `init` | `--assistants <ids>`, `--no-interactive`, `--apply`, `--no-diff`, `--json`, `--help` |
-| `detect` | `--json`, `--help` |
+| `init` | `--assistants <ids>`, `--provider <id>`, `--no-interactive`, `--apply`, `--no-diff`, `--json`, `--help` |
+| `detect` | `--provider <id>`, `--json`, `--help` |
+| `run` | `--provider <id>`, right before or after the assistant |
 | `undo` | `--no-interactive`, `--apply`, `--no-diff`, `--json`, `--list`, `--help` |
 
 `apex init --wat` fails fast, and so does a flag in the wrong command
 (`apex detect --no-diff`). Only `--apply` writes, so there is no "preview *and* write" to argue
 about. `--assistants <ids>` is a comma list of `opencode`, `codex`,
-`claude`, `pi`, `cursor`, `copilot`, `ai-sdk`.
+`claude`, `pi`, `cursor`, `copilot`, `ai-sdk`. `--provider <id>` is `vercel` or `callstack`.
 
 Non-interactive callers (`--no-interactive`, `--json`, pipes, CI) never get a prompt and never
 have files changed unless they also pass `--apply`. `--json` prints a machine-readable plan with
-`mode`, `applied`, `appliedPaths`, per-file `changes` and `nextSteps` instead of the human UI. If a
+`mode`, `provider`, `model`, `envVar`, `applied`, `appliedPaths`, per-file `changes` and `nextSteps` instead of the human UI. If a
 write fails halfway through a batch, the payload still reports what already landed plus the `error`,
 and the exit code is 1.
 
@@ -64,15 +80,15 @@ apex completion fish > ~/.config/fish/completions/apex.fish
 
 ```
  ┌────────────────────────────────────────────────────────────────────────────┐
- │ Apex CLI · Configure callstack/Apex for your favorite harness              │
+ │ Apex CLI · Configure Apex for your favorite harness                        │
  │                                                                            │
  │ Nothing is written until you confirm.                                      │
  └────────────────────────────────────────────────────────────────────────────┘
 │
-◆  Set up callstack/Apex for which assistants?  (space toggles, enter confirms)
-│  OpenCode, Codex
+◇  Which gateway should your assistants use for Apex?
+│  Vercel AI Gateway
 │
-◇  Set up callstack/Apex for which assistants?  (space toggles, enter confirms)
+◇  Set up callstack/apex for which assistants?  (space toggles, enter confirms)
 │  OpenCode, Codex
 
 
@@ -87,15 +103,15 @@ apex completion fish > ~/.config/fish/completions/apex.fish
    -  "theme": "dark"
    +  "theme": "dark",
    +  "provider": {
-   +    "callstack.ai": { … }
+   +    "apex-vercel": { … }
    +  }
     }
 
-   Create     ~/.codex/callstack_ai.config.toml
+   Create     ~/.codex/apex_vercel.config.toml
    --- /dev/null
-   +++ ~/.codex/callstack_ai.config.toml
+   +++ ~/.codex/apex_vercel.config.toml
    @@ -0,0 +1,10 @@
-   +model_provider = "callstack_ai"
+   +model_provider = "apex_vercel"
 
    Existing files get a .apex-backup-<id> copy first.        (dimmed)
 │
@@ -105,24 +121,25 @@ apex completion fish > ~/.config/fish/completions/apex.fish
 
    Environment
 
-   ✓ CALLSTACK_AUTH_TOKEN is set in this shell
+   ✓ AI_GATEWAY_API_KEY is set in this shell
    Apex CLI never reads, stores or logs the key itself.
 
 
-   Use these commands to run callstack/Apex with your selected harnesses:
+   Use these commands to run callstack/apex with your selected harnesses:
 
-   apex run opencode  opencode --model callstack.ai/callstack/Apex        (dimmed)
-   apex run codex     codex --profile callstack_ai
+   apex run opencode  opencode --model apex-vercel/callstack/apex         (dimmed)
+   apex run codex     codex --profile apex_vercel
 
-   ...or pick "callstack/Apex" from the UI when setting up manually.
-   For more instructions, visit:
-   https://app.notion.com/p/callstack/Apex-how-to-use-it-36d5d027c0f880e99d03d1c37a77382f
+   ...or pick "callstack/apex" from the UI when setting up manually.
+   For more instructions, visit: https://vercel.com/docs/ai-gateway/coding-agents
 
    If you want to undo the changes, run apex undo
 ```
-`Environment` only shows the `export CALLSTACK_AUTH_TOKEN=…` line while the variable is missing; once
-it is set you get a green `✓ CALLSTACK_AUTH_TOKEN is set in this shell` instead. The closing block shows
-what each `apex run` expands to, links the guide (the URL is never folded, so it stays clickable), and
+`Environment` only shows the `export AI_GATEWAY_API_KEY=…` line (and, for Vercel AI Gateway, where to
+create a key) while the variable is missing; once it is set you get a green
+`✓ AI_GATEWAY_API_KEY is set in this shell` instead. The Callstack portal shows `CALLSTACK_AUTH_TOKEN`. The closing block shows
+what each `apex run` expands to, links the gateway's guide (Vercel's coding-agents docs or the Callstack portal guide; the URL is never
+folded, so it stays clickable), and
 mentions `apex undo` only when something was actually written.
 
 The diff is the change list: it only shows the lines that actually move, paths are shortened to
@@ -143,15 +160,17 @@ their own backup first, so an undo is itself reversible. Apex CLI records each b
 
 ## Integrations
 
+Names below are Vercel AI Gateway / Callstack portal.
+
 | Assistant | Configuration | Behavior |
 | --- | --- | --- |
-| OpenCode | `~/.config/opencode/opencode.json` or `.jsonc` | Adds `provider.callstack.ai`, OpenAI-compatible transport, model and environment-key reference; preserves the default model. A config already in OpenCode 2's `providers` format gets the v2 entry instead, and you add the key with `/connect`. |
-| Codex 0.134.0+ | `~/.codex/callstack_ai.config.toml` | Adds a self-contained Responses API provider/profile; leaves base config and default model untouched. |
-| Claude Code | `~/.claude/settings.json` | Merges attribution flag only; use `apex run claude` for gateway credentials and model selection. |
-| Pi | `~/.pi/agent/models.json` and `settings.json` | Adds `providers.callstack` with Chat Completions, Apex and `$CALLSTACK_AUTH_TOKEN`; saves `medium` as Apex's per-model thinking default for direct launches. Preserves other models and their defaults. Requires a Pi version supporting `$VAR` key interpolation and `modelThinkingLevels`. |
-| Cursor | Guided setup | Prints endpoint, API-key and custom-model steps; does not modify private editor storage. |
+| OpenCode | `~/.config/opencode/opencode.json` or `.jsonc` | Adds `provider.apex-vercel` / `provider.callstack.ai`, OpenAI-compatible transport, model and environment-key reference; preserves the default model. A config already in OpenCode 2's `providers` format gets the v2 entry instead, and you add the key with `/connect`. |
+| Codex 0.134.0+ | `~/.codex/apex_vercel.config.toml` / `callstack_ai.config.toml` | Adds a self-contained Responses API provider/profile (Vercel's Codex endpoint `/codex/v1`); leaves base config and default model untouched. |
+| Claude Code | `~/.claude/settings.json` | Merges attribution flag only; use `apex run claude` for gateway credentials and model selection (Vercel's Claude Code endpoint `/claude-code`). |
+| Pi | `~/.pi/agent/models.json` and `settings.json` | Adds `providers.apex-vercel` / `providers.callstack` with Chat Completions, Apex and `$AI_GATEWAY_API_KEY` / `$CALLSTACK_AUTH_TOKEN`; saves `medium` as Apex's per-model thinking default for direct launches. Preserves other models and their defaults. Requires a Pi version supporting `$VAR` key interpolation and `modelThinkingLevels`. |
+| Cursor | Guided setup | Prints endpoint (Vercel's Cursor endpoint `/cursor/v1`), API-key and custom-model steps; does not modify private editor storage. |
 | VS Code / Copilot | Guided setup | Detects VS Code, not whether Copilot is installed; prints custom-endpoint steps and model JSON, retaining the editor-generated secret reference. |
-| Vercel AI SDK / Eve | Guided setup | Detected from the `package.json` in the directory you run Apex CLI from (`ai`, `@ai-sdk/openai` or `eve`); prints the connector snippet, which reads the key from `CALLSTACK_AUTH_TOKEN`. Nothing is written into the project. |
+| Vercel AI SDK / Eve | Guided setup | Detected from the `package.json` in the directory you run Apex CLI from (`ai`, `@ai-sdk/openai` or `eve`); prints the connector snippet, which reads the key from `AI_GATEWAY_API_KEY` or `CALLSTACK_AUTH_TOKEN` (Vercel AI Gateway uses its general `/v1` endpoint here, since an app is not a coding agent). Nothing is written into the project. |
 
 Apex configs declare tool calling, image input, and the supported
 `none`/`low`/`medium`/`xhigh` reasoning efforts (default `medium`) where supported.
@@ -163,8 +182,8 @@ reserve.
 Pi's per-model default applies to fresh direct launches and model selection; `apex run pi` also
 selects `medium` explicitly. Resumed Pi sessions retain their saved thinking level.
 
-The gateway model ID, authentication, `medium` default, and `off` → `none` translation remain
-Callstack-specific.
+The model ID and key depend on the gateway you pick; the `medium` default and `off` → `none`
+translation are Apex-specific.
 
 | Harness | Context field | Output field | Official reference |
 | --- | --- | --- | --- |
@@ -197,16 +216,20 @@ avoid shell interpolation; launch the configured assistant directly or use WSL.
 
 ## Authentication
 
-Provide `CALLSTACK_AUTH_TOKEN` through your shell or a secret manager. Apex CLI never prompts for,
-logs, stores or validates the key, and never writes it to a config file; configs only reference it.
-A hidden prompt keeps it out of your shell history:
+Provide `AI_GATEWAY_API_KEY` (Vercel AI Gateway) or `CALLSTACK_AUTH_TOKEN` (Callstack portal)
+through your shell or a secret manager. Create a Vercel key on the AI Gateway page of your Vercel
+dashboard. Apex CLI never prompts for, logs, stores or validates the key, and never writes it to a
+config file; configs only reference it. A hidden prompt keeps it out of your shell history:
 
 ```bash
-read -r -s -p 'Callstack API key: ' CALLSTACK_AUTH_TOKEN; echo
-export CALLSTACK_AUTH_TOKEN
+read -r -s -p 'AI Gateway API key: ' AI_GATEWAY_API_KEY; echo
+export AI_GATEWAY_API_KEY
 apex run codex
 ```
 
+With both keys exported, `apex run` uses Vercel AI Gateway; pass `--provider callstack` for the
+portal: `apex run --provider callstack codex`.
+
 Other launchers: `apex run opencode`, `apex run claude`, `apex run pi`, `apex run codex -- --help`.
 Launchers pass extra arguments through unchanged (without a shell), can be overridden by explicit
 user arguments, and never run `init`. `apex run claude` sets the gateway URL, token, attribution
@@ -221,7 +244,7 @@ Apex CLI will not break what you already have.
 - It copies a file next to itself before it edits that file, so you can always put the old one back.
 - It keeps your comments, your other settings, and your default model.
 - It never asks for your API key, writes it down, or sends it anywhere. You keep it in
-  `CALLSTACK_AUTH_TOKEN`, and the CLI prints `<redacted>` wherever a secret would show.
+  `AI_GATEWAY_API_KEY` or `CALLSTACK_AUTH_TOKEN`, and the CLI prints `<redacted>` wherever a secret would show.
 - If a config file is broken, set up another way, or a link to somewhere else, Apex CLI stops and
   tells you. It does not overwrite it.
 - `apex undo` gives your old setup back. It only touches files that still match what Apex CLI wrote.
@@ -235,10 +258,10 @@ How that works underneath:
   then land in your dotfiles repo, so add `*.apex-backup-*` to its `.gitignore`.
 - The dedicated Codex Apex profile's owned keys are upgraded rather than rejected as conflicts;
   the base `config.toml` and other profiles remain untouched.
-- JSON/JSONC edits preserve comments and unrelated keys. The Callstack endpoint is set to the
+- JSON/JSONC edits preserve comments and unrelated keys. The gateway endpoint is set to the
   documented one. A key the assistant already has (in OpenCode's own key store, or in Pi's
   `models.json`, where Apex CLI 0.2 put them) is kept, so upgrading never leaves an assistant
-  without a key; otherwise the config references `CALLSTACK_AUTH_TOKEN`.
+  without a key; otherwise the config references `AI_GATEWAY_API_KEY` or `CALLSTACK_AUTH_TOKEN`.
 - Writes use same-directory temporary files plus rename with mode `0600` (new directories `0700`),
   and abort if the file changed on disk between planning and saving. This is not a lock or a
   multi-file transaction: a later I/O failure can leave earlier writes applied. `apex undo`
@@ -255,6 +278,8 @@ How that works underneath:
   (separate profile files for 0.134.0+; provider is placed in that profile).
 - OpenCode v1 configuration: `https://dev.opencode.ai/docs/config/`
   (JSONC and `{env:VARIABLE}` support).
+- Vercel AI Gateway for coding agents: `https://vercel.com/docs/ai-gateway/coding-agents`
+  (per-agent endpoints; Apex is listed as `callstack/apex`).
 - Pi custom models: `https://github.com/badlogic/pi-mono/blob/main/packages/coding-agent/docs/models.md`
   (`$VARIABLE` key interpolation; bare uppercase names are literals in current docs).
 

package/dist/cli.js

@@ -2451,9 +2451,50 @@ async function writeChange(change) {
 
 //#endregion
 //#region src/assistants.js
-const MODEL = "callstack/Apex";
-const BASE_URL = "https://api.callstack.ai/v1";
-const GUIDE_URL = "https://app.notion.com/p/callstack/Apex-how-to-use-it-36d5d027c0f880e99d03d1c37a77382f";
+const VERCEL = "https://ai-gateway.vercel.sh";
+const PROVIDERS = {
+	vercel: {
+		id: "vercel",
+		name: "Vercel AI Gateway",
+		label: "Vercel AI Gateway",
+		model: "callstack/apex",
+		envVar: "AI_GATEWAY_API_KEY",
+		keyName: "your Vercel AI Gateway API key",
+		keysUrl: "https://vercel.com/d?to=%2F%5Bteam%5D%2F%7E%2Fai-gateway&title=AI+Gateway",
+		guideUrl: "https://vercel.com/docs/ai-gateway/coding-agents",
+		baseUrl: `${VERCEL}/coding-agent/v1`,
+		codexUrl: `${VERCEL}/codex/v1`,
+		cursorUrl: `${VERCEL}/cursor/v1`,
+		anthropicUrl: `${VERCEL}/claude-code`,
+		appUrl: `${VERCEL}/v1`,
+		opencode: "apex-vercel",
+		pi: "apex-vercel",
+		codex: "apex_vercel"
+	},
+	callstack: {
+		id: "callstack",
+		name: "Callstack portal",
+		label: "callstack.ai",
+		model: "callstack/Apex",
+		envVar: "CALLSTACK_AUTH_TOKEN",
+		keyName: "your callstack.ai key",
+		keysUrl: null,
+		guideUrl: "https://app.notion.com/p/callstack/Apex-how-to-use-it-36d5d027c0f880e99d03d1c37a77382f",
+		baseUrl: "https://api.callstack.ai/v1",
+		codexUrl: "https://api.callstack.ai/v1",
+		cursorUrl: "https://api.callstack.ai/v1",
+		anthropicUrl: "https://api.callstack.ai",
+		appUrl: "https://api.callstack.ai/v1",
+		opencode: "callstack.ai",
+		pi: "callstack",
+		codex: "callstack_ai"
+	}
+};
+const PROVIDER_IDS = Object.keys(PROVIDERS);
+function defaultProvider(env = process.env) {
+	const keyed = PROVIDER_IDS.filter((id) => env[PROVIDERS[id].envVar]?.trim());
+	return PROVIDERS[keyed.length === 1 ? keyed[0] : "vercel"];
+}
 const IDS = [
 	"opencode",
 	"codex",
@@ -2502,30 +2543,25 @@ const OPENCODE_MODEL = {
 	options: { reasoningEffort: "medium" },
 	variants: Object.fromEntries(EFFORTS.map((effort) => [effort, { reasoningEffort: effort }]))
 };
-const OPENCODE_V2_PROVIDER = {
-	name: "callstack.ai",
-	package: "@opencode/ai/providers/openai-compatible",
-	settings: { baseURL: BASE_URL },
-	models: { [MODEL]: {
-		name: "Apex",
-		capabilities: {
-			tools: true,
-			input: ["text", "image"],
-			output: ["text"]
-		},
-		limit: {
-			context: CONTEXT_WINDOW,
-			output: MAX_OUTPUT
-		},
-		settings: { reasoningEffort: "medium" },
-		variants: EFFORTS.map((effort) => ({
-			id: effort,
-			settings: { reasoningEffort: effort }
-		}))
-	} }
+const OPENCODE_V2_PACKAGE = "@opencode/ai/providers/openai-compatible";
+const OPENCODE_V2_MODEL = {
+	name: "Apex",
+	capabilities: {
+		tools: true,
+		input: ["text", "image"],
+		output: ["text"]
+	},
+	limit: {
+		context: CONTEXT_WINDOW,
+		output: MAX_OUTPUT
+	},
+	settings: { reasoningEffort: "medium" },
+	variants: EFFORTS.map((effort) => ({
+		id: effort,
+		settings: { reasoningEffort: effort }
+	}))
 };
 const PI_MODEL = {
-	id: MODEL,
 	reasoning: true,
 	input: ["text", "image"],
 	thinkingLevelMap: {
@@ -2632,16 +2668,17 @@ async function detect(options = {}) {
 		};
 	}));
 }
-async function openCodeHasKey(authFile) {
+async function openCodeHasKey(authFile, providerId) {
 	if (!authFile) return false;
 	try {
-		return Object.hasOwn(JSON.parse(await readConfig(authFile) ?? "{}"), "callstack.ai");
+		return Object.hasOwn(JSON.parse(await readConfig(authFile) ?? "{}"), providerId);
 	} catch {
 		return false;
 	}
 }
 const PI_PLACEHOLDER = "XXX";
-async function planAssistant(assistant) {
+async function planAssistant(assistant, provider = PROVIDERS.vercel) {
+	const { model, envVar } = provider;
 	const changes = [];
 	const json = async (filename, updates) => {
 		const path = join(assistant.directory, filename);
@@ -2660,124 +2697,131 @@ async function planAssistant(assistant) {
 			const hasJsonc = await exists(join(assistant.directory, "opencode.jsonc"));
 			if (hasJson && hasJsonc) throw new Error("Both opencode.json and opencode.jsonc exist. Consolidate them before init.");
 			let v2 = false;
-			const storedKey = await openCodeHasKey(assistant.authFile);
+			const id = provider.opencode;
+			const storedKey = await openCodeHasKey(assistant.authFile, id);
 			await json(hasJsonc ? "opencode.jsonc" : "opencode.json", (existing) => {
 				v2 = Object.hasOwn(existing, "providers");
 				if (v2) return [
 					[[
 						"providers",
-						"callstack.ai",
+						id,
 						"name"
-					], OPENCODE_V2_PROVIDER.name],
+					], provider.label],
 					[[
 						"providers",
-						"callstack.ai",
+						id,
 						"package"
-					], OPENCODE_V2_PROVIDER.package],
+					], OPENCODE_V2_PACKAGE],
 					[[
 						"providers",
-						"callstack.ai",
+						id,
 						"settings",
 						"baseURL"
-					], BASE_URL],
-					...Object.entries(OPENCODE_V2_PROVIDER.models[MODEL]).map(([key, value]) => [[
+					], provider.baseUrl],
+					...Object.entries(OPENCODE_V2_MODEL).map(([key, value]) => [[
 						"providers",
-						"callstack.ai",
+						id,
 						"models",
-						MODEL,
+						model,
 						key
 					], key === "settings" ? {
-						...existing.providers?.["callstack.ai"]?.models?.[MODEL]?.settings,
+						...existing.providers?.[id]?.models?.[model]?.settings,
 						...value
 					} : value])
 				];
 				return [
 					[[
 						"provider",
-						"callstack.ai",
+						id,
 						"npm"
 					], "@ai-sdk/openai-compatible"],
 					[[
 						"provider",
-						"callstack.ai",
+						id,
 						"name"
-					], "callstack.ai"],
+					], provider.label],
 					[[
 						"provider",
-						"callstack.ai",
+						id,
 						"options",
 						"baseURL"
-					], BASE_URL],
-					...storedKey || existing.provider?.["callstack.ai"]?.options?.apiKey ? [] : [[[
+					], provider.baseUrl],
+					...storedKey || existing.provider?.[id]?.options?.apiKey ? [] : [[[
 						"provider",
-						"callstack.ai",
+						id,
 						"options",
 						"apiKey"
-					], "{env:CALLSTACK_AUTH_TOKEN}"]],
+					], `{env:${envVar}}`]],
 					...Object.entries(OPENCODE_MODEL).map(([key, value]) => [[
 						"provider",
-						"callstack.ai",
+						id,
 						"models",
-						MODEL,
+						model,
 						key
 					], key === "options" ? {
-						...existing.provider?.["callstack.ai"]?.models?.[MODEL]?.options,
+						...existing.provider?.[id]?.models?.[model]?.options,
 						...value
 					} : value])
 				];
 			});
-			if (v2) changes[0].steps = ["OpenCode 2 stores provider keys itself: run /connect in OpenCode, pick callstack.ai → \"Manually enter API Key\", and paste your Callstack key."];
+			if (v2) changes[0].steps = [`OpenCode 2 stores provider keys itself: run /connect in OpenCode, pick ${provider.label} → "Manually enter API Key", and paste ${provider.keyName}.`];
 			break;
 		}
-		case "pi":
+		case "pi": {
+			const id = provider.pi;
+			const entry = {
+				id: model,
+				...PI_MODEL
+			};
 			await json("models.json", (existing) => {
-				const models = existing.providers?.callstack?.models ?? [];
-				const key = existing.providers?.callstack?.apiKey;
-				if (!Array.isArray(models) || models.some((model) => !model || typeof model.id !== "string")) throw new Error("Expected Pi models to be an array of model objects.");
+				const models = existing.providers?.[id]?.models ?? [];
+				const key = existing.providers?.[id]?.apiKey;
+				if (!Array.isArray(models) || models.some((item) => !item || typeof item.id !== "string")) throw new Error("Expected Pi models to be an array of model objects.");
 				return [
 					[[
 						"providers",
-						"callstack",
+						id,
 						"baseUrl"
-					], BASE_URL],
+					], provider.baseUrl],
 					[[
 						"providers",
-						"callstack",
+						id,
 						"api"
 					], "openai-completions"],
 					...key && key !== PI_PLACEHOLDER ? [] : [[[
 						"providers",
-						"callstack",
+						id,
 						"apiKey"
-					], "$CALLSTACK_AUTH_TOKEN"]],
+					], `$${envVar}`]],
 					[[
 						"providers",
-						"callstack",
+						id,
 						"models"
-					], models.some((model) => model.id === "callstack/Apex") ? models.map((model) => model.id === "callstack/Apex" ? {
-						...model,
-						...PI_MODEL,
-						...model.compat ? { compat: Object.fromEntries(Object.entries(model.compat).filter(([key]) => key !== "reasoningEffortMap")) } : {}
-					} : model) : [...models, PI_MODEL]]
+					], models.some((item) => item.id === model) ? models.map((item) => item.id === model ? {
+						...item,
+						...entry,
+						...item.compat ? { compat: Object.fromEntries(Object.entries(item.compat).filter(([key]) => key !== "reasoningEffortMap")) } : {}
+					} : item) : [...models, entry]]
 				];
 			});
-			await json("settings.json", [[["modelThinkingLevels", `callstack/${MODEL}`], "medium"]]);
+			await json("settings.json", [[["modelThinkingLevels", `${id}/${model}`], "medium"]]);
 			break;
+		}
 		case "claude":
 			await json("settings.json", [[["env", "CLAUDE_CODE_ATTRIBUTION_HEADER"], "0"]]);
 			break;
 		case "codex": {
-			const path = join(assistant.directory, "callstack_ai.config.toml");
+			const path = join(assistant.directory, `${provider.codex}.config.toml`);
 			const before = await readConfig(path);
 			const after = updateToml(before, {
-				model_provider: "callstack_ai",
-				model: MODEL,
+				model_provider: provider.codex,
+				model,
 				model_context_window: CONTEXT_WINDOW,
 				model_reasoning_effort: "medium",
-				model_providers: { callstack_ai: {
-					name: "callstack.ai",
-					base_url: BASE_URL,
-					env_key: "CALLSTACK_AUTH_TOKEN",
+				model_providers: { [provider.codex]: {
+					name: provider.label,
+					base_url: provider.codexUrl,
+					env_key: envVar,
 					wire_api: "responses",
 					requires_openai_auth: false
 				} }
@@ -2793,25 +2837,26 @@ async function planAssistant(assistant) {
 	}
 	return changes;
 }
-function launchOptions(id, env = process.env) {
-	const token = env.CALLSTACK_AUTH_TOKEN;
-	if (!token?.trim()) throw new Error("Set CALLSTACK_AUTH_TOKEN in your environment before launching.");
+function launchOptions(id, env = process.env, provider = defaultProvider(env)) {
+	const { model, envVar } = provider;
+	const token = env[envVar];
+	if (!token?.trim()) throw new Error(`Set ${envVar} in your environment before launching with ${provider.name}.`);
 	const childEnv = { ...env };
 	switch (id) {
 		case "opencode": return {
-			args: ["--model", `callstack.ai/${MODEL}`],
+			args: ["--model", `${provider.opencode}/${model}`],
 			env: childEnv
 		};
 		case "codex": return {
-			args: ["--profile", "callstack_ai"],
+			args: ["--profile", provider.codex],
 			env: childEnv
 		};
 		case "pi": return {
 			args: [
 				"--provider",
-				"callstack",
+				provider.pi,
 				"--model",
-				MODEL,
+				model,
 				"--thinking",
 				"medium"
 			],
@@ -2826,33 +2871,33 @@ function launchOptions(id, env = process.env) {
 				"CLAUDE_CODE_USE_FOUNDRY"
 			]) delete childEnv[key];
 			Object.assign(childEnv, {
-				ANTHROPIC_BASE_URL: "https://api.callstack.ai",
+				ANTHROPIC_BASE_URL: provider.anthropicUrl,
 				ANTHROPIC_AUTH_TOKEN: token,
-				ANTHROPIC_MODEL: MODEL,
+				ANTHROPIC_MODEL: model,
 				CLAUDE_CODE_ATTRIBUTION_HEADER: "0",
-				ANTHROPIC_DEFAULT_OPUS_MODEL: MODEL,
-				ANTHROPIC_DEFAULT_SONNET_MODEL: MODEL,
-				ANTHROPIC_DEFAULT_HAIKU_MODEL: MODEL,
-				CLAUDE_CODE_SUBAGENT_MODEL: MODEL,
+				ANTHROPIC_DEFAULT_OPUS_MODEL: model,
+				ANTHROPIC_DEFAULT_SONNET_MODEL: model,
+				ANTHROPIC_DEFAULT_HAIKU_MODEL: model,
+				CLAUDE_CODE_SUBAGENT_MODEL: model,
 				CLAUDE_CODE_MAX_CONTEXT_TOKENS: String(CONTEXT_WINDOW),
 				CLAUDE_CODE_MAX_OUTPUT_TOKENS: String(MAX_OUTPUT)
 			});
 			return {
-				args: ["--model", MODEL],
+				args: ["--model", model],
 				env: childEnv
 			};
 		default: throw new Error(`Cannot launch ${id}. Use the editor's model selector.`);
 	}
 }
 const RUN_NOTE = { claude: "(ANTHROPIC_* gateway environment set)" };
-function runExpansion(id) {
-	const { args } = launchOptions(id, { CALLSTACK_AUTH_TOKEN: "placeholder" });
+function runExpansion(id, provider) {
+	const { args } = launchOptions(id, { [provider.envVar]: "placeholder" }, provider);
 	return [`${id} ${args.join(" ")}`, RUN_NOTE[id]].filter(Boolean).join("  ");
 }
-const COPILOT_MODEL = JSON.stringify({
-	id: MODEL,
+const copilotModel = (provider) => JSON.stringify({
+	id: provider.model,
 	name: "Apex",
-	url: BASE_URL,
+	url: provider.baseUrl,
 	toolCalling: true,
 	vision: true,
 	thinking: true,
@@ -2861,32 +2906,39 @@ const COPILOT_MODEL = JSON.stringify({
 	supportsReasoningEffort: EFFORTS,
 	reasoningEffortFormat: "chat-completions"
 }, null, 2);
-const AI_SDK_SNIPPET = [
+const aiSdkSnippet = (provider) => [
 	"import { createOpenAI } from '@ai-sdk/openai';",
 	"",
 	"const apex = createOpenAI({",
 	"  name: 'callstack',",
-	`  baseURL: '${BASE_URL}',`,
-	"  apiKey: process.env.CALLSTACK_AUTH_TOKEN,",
+	`  baseURL: '${provider.appUrl}',`,
+	`  apiKey: process.env.${provider.envVar},`,
 	"});",
 	"const requestOptions = {",
-	`  model: apex('${MODEL}'),`,
+	`  model: apex('${provider.model}'),`,
 	`  maxOutputTokens: ${MAX_OUTPUT},`,
 	"};",
 	"// Pass requestOptions with your prompt to generateText / streamText.",
 	`// Reasoning effort (${EFFORTS.join(", ")}):`,
 	"// providerOptions: { callstack: { reasoningEffort: 'medium' } }"
 ];
-const MANUAL = {
-	cursor: [`Cursor: Settings → Models → API Keys → OpenAI API Key. Enter your Callstack key, override the base URL with ${BASE_URL}, add and enable ${MODEL}, then select it in Agent.`, "Cursor has no way to declare what a custom model can do (tool calling, vision, reasoning); it infers that itself."],
-	copilot: [
-		"Installation of Copilot itself is not verified: Copilot → model selector → Manage Models → Add Models → Custom Endpoint → name callstack.ai → enter your key → Chat Completions.",
-		"Keep the generated apiKey secret reference; add this object to its models array:",
-		...COPILOT_MODEL.split("\n")
-	],
-	"ai-sdk": ["Apex CLI never writes your key into a project. Keep it in CALLSTACK_AUTH_TOKEN (for example in a .env that git ignores) and create the provider with:", ...AI_SDK_SNIPPET]
-};
-const MANUAL_IDS = Object.keys(MANUAL);
+const MANUAL_IDS = [
+	"cursor",
+	"copilot",
+	"ai-sdk"
+];
+function manualSteps(id, provider) {
+	switch (id) {
+		case "cursor": return [`Cursor: Settings → Models → API Keys → OpenAI API Key. Enter ${provider.keyName}, override the base URL with ${provider.cursorUrl}, add and enable ${provider.model}, then select it in Agent.`, "Cursor has no way to declare what a custom model can do (tool calling, vision, reasoning); it infers that itself."];
+		case "copilot": return [
+			`Installation of Copilot itself is not verified: Copilot → model selector → Manage Models → Add Models → Custom Endpoint → name ${provider.label} → enter ${provider.keyName} → Chat Completions.`,
+			"Keep the generated apiKey secret reference; add this object to its models array:",
+			...copilotModel(provider).split("\n")
+		];
+		case "ai-sdk": return [`Apex CLI never writes your key into a project. Keep it in ${provider.envVar} (for example in a .env that git ignores) and create the provider with:`, ...aiSdkSnippet(provider)];
+		default: throw new Error(`${id} is set up automatically.`);
+	}
+}
 
 //#endregion
 //#region node_modules/fast-string-truncated-width/dist/utils.js
@@ -3585,6 +3637,32 @@ var a = class extends y {
 		});
 	}
 };
+let n$1 = class n extends y {
+	options;
+	cursor = 0;
+	get _selectedValue() {
+		return this.options[this.cursor];
+	}
+	changeValue() {
+		const e = this._selectedValue;
+		this.value = e === void 0 ? void 0 : e.value;
+	}
+	constructor(e) {
+		super(e, false), this.options = e.options;
+		const o = this.options.findIndex(({ value: s }) => s === e.initialValue), t = o === -1 ? 0 : o;
+		this.cursor = this.options[t]?.disabled ? findCursor(t, 1, this.options) : t, this.changeValue(), this.on("cursor", (s) => {
+			switch (s) {
+				case "left":
+				case "up":
+					this.cursor = findCursor(this.cursor, -1, this.options);
+					break;
+				case "down":
+				case "right": this.cursor = findCursor(this.cursor, 1, this.options);
+			}
+			this.changeValue();
+		});
+	}
+};
 
 //#endregion
 //#region node_modules/@clack/prompts/dist/index.mjs
@@ -3816,6 +3894,63 @@ const u = {
 	block: unicodeOr("█", "#")
 };
 const SELECT_INSTRUCTIONS = [`${styleText("dim", "↑/↓")} to navigate`, `${styleText("dim", "Enter:")} confirm`];
+const c = (t, o) => t.includes(`
+`) ? t.split(`
+`).map((d) => o(d)).join(`
+`) : o(t);
+const select = (t) => {
+	const o = (n, m) => {
+		if (n === void 0) return "";
+		const s = n.label ?? String(n.value);
+		switch (m) {
+			case "disabled": return `${styleText("gray", S_RADIO_INACTIVE)} ${c(s, (i) => styleText("gray", i))}${n.hint ? ` ${styleText("dim", `(${n.hint ?? "disabled"})`)}` : ""}`;
+			case "selected": return `${c(s, (i) => styleText("dim", i))}`;
+			case "active": return `${styleText("green", S_RADIO_ACTIVE)} ${s}${n.hint ? ` ${styleText("dim", `(${n.hint})`)}` : ""}`;
+			case "cancelled": return `${c(s, (i) => styleText(["strikethrough", "dim"], i))}`;
+			default: return `${styleText("dim", S_RADIO_INACTIVE)} ${c(s, (i) => styleText("dim", i))}`;
+		}
+	}, d = t.showInstructions ?? true;
+	return new n$1({
+		options: t.options,
+		signal: t.signal,
+		input: t.input,
+		output: t.output,
+		initialValue: t.initialValue,
+		render() {
+			const n = t.withGuide ?? settings.withGuide, m = `${symbol(this.state)}  `, s = `${symbolBar(this.state)}  `, i = wrapTextWithPrefix(t.output, t.message, s, m), u = `${n ? `${styleText("gray", S_BAR)}
+` : ""}${i}
+`;
+			switch (this.state) {
+				case "submit": {
+					const r = n ? `${styleText("gray", S_BAR)}  ` : "";
+					return `${u}${wrapTextWithPrefix(t.output, o(this.options[this.cursor], "selected"), r)}`;
+				}
+				case "cancel": {
+					const r = n ? `${styleText("gray", S_BAR)}  ` : "";
+					return `${u}${wrapTextWithPrefix(t.output, o(this.options[this.cursor], "cancelled"), r)}${n ? `
+${styleText("gray", S_BAR)}` : ""}`;
+				}
+				default: {
+					const r = n ? `${styleText("cyan", S_BAR)}  ` : "", a = u.split(`
+`).length, p = d ? formatInstructionFooter(SELECT_INSTRUCTIONS, n) : n ? [styleText("cyan", S_BAR_END)] : [], f = p.join(`
+`), b = p.length + 1;
+					return `${u}${r}${limitOptions({
+						output: t.output,
+						cursor: this.cursor,
+						options: this.options,
+						maxItems: t.maxItems,
+						columnPadding: r.length,
+						rowPadding: a + b,
+						style: (g, x) => o(g, g.disabled ? "disabled" : x ? "active" : "inactive")
+					}).join(`
+${r}`)}
+${f}
+`;
+				}
+			}
+		}
+	}).prompt();
+};
 const i = `${styleText("gray", S_BAR)}  `;
 
 //#endregion
@@ -4072,7 +4207,8 @@ const pad = (text, column) => {
 	const value = String(text);
 	return value + " ".repeat(Math.max(1, column - visible(value).length));
 };
-const MODEL_PATTERN = new RegExp(`(?<![\\w/.-])${MODEL.replace(/[.*+?^${}()|[\]\\]/g, "\\$&")}(?![\\w-])`, "g");
+const MODELS = [...new Set(Object.values(PROVIDERS).map((provider) => provider.model))].map((id) => id.replace(/[.*+?^${}()|[\]\\]/g, "\\$&"));
+const MODEL_PATTERN = new RegExp(`(?<![\\w/.-])(?:${MODELS.join("|")})(?![\\w-])`, "g");
 const model = (text) => String(text).replace(MODEL_PATTERN, (match) => green(match));
 const STATUS_COLUMN = columnWidth(Object.values(STATUS).map((status) => status.word));
 function clip(text, limit) {
@@ -4176,9 +4312,21 @@ function bullet(marker, message) {
 const success = (message) => bullet(green("✔"), message);
 const warn = (message) => bullet(yellow("!"), message);
 const error = (message) => bullet(red("✖"), message);
-async function selectAssistants(options, initial) {
+async function selectProvider(options, initial) {
+	const chosen = await select({
+		message: "Which gateway should your assistants use for Apex?",
+		options,
+		initialValue: initial
+	});
+	if (isCancel(chosen)) {
+		cancel("Cancelled. Nothing was written.");
+		return false;
+	}
+	return chosen;
+}
+async function selectAssistants(options, initial, modelId) {
 	const chosen = await multiselect({
-		message: model(`Set up ${MODEL} for which assistants?  ${dim("(space toggles, enter confirms)")}`),
+		message: model(`Set up ${modelId} for which assistants?  ${dim("(space toggles, enter confirms)")}`),
 		options,
 		initialValues: initial,
 		required: false,
@@ -4387,6 +4535,12 @@ const FLAGS = {
 		hint: "assistant ids",
 		text: `comma list of: ${IDS.join(", ")} (skips the picker)`
 	},
+	provider: {
+		type: "string",
+		usage: "--provider <id>",
+		hint: "gateway serving Apex",
+		text: `${PROVIDER_IDS.join(" or ")}; defaults to the one whose key is set, else vercel (skips the question)`
+	},
 	"no-interactive": {
 		type: "boolean",
 		usage: "--no-interactive",
@@ -4431,6 +4585,7 @@ const COMMANDS = {
 		summary: "configure assistants",
 		flags: [
 			"assistants",
+			"provider",
 			"no-interactive",
 			"apply",
 			"no-diff",
@@ -4439,9 +4594,13 @@ const COMMANDS = {
 		]
 	},
 	detect: {
-		args: "[--json]",
+		args: "[--provider <id>] [--json]",
 		summary: "show what Apex CLI can see on this machine",
-		flags: ["json", "help"]
+		flags: [
+			"provider",
+			"json",
+			"help"
+		]
 	},
 	undo: {
 		args: "[--list] [options]",
@@ -4456,7 +4615,7 @@ const COMMANDS = {
 		]
 	},
 	run: {
-		args: `<${RUNNABLE.join("|")}> [-- <args>]`,
+		args: `[--provider <id>] <${RUNNABLE.join("|")}> [-- <args>]`,
 		summary: "launch an assistant with the gateway set",
 		flags: []
 	},
@@ -4520,12 +4679,13 @@ const exampleRows = [
 	["apex init", "pick assistants, review, confirm"],
 	["apex init --assistants codex,pi", "preview two tools (agent and CI friendly)"],
 	["apex init --assistants codex,pi --apply", "write those two without asking"],
+	["apex init --provider callstack", "use the Callstack portal instead of Vercel AI Gateway"],
 	["apex undo", "preview the reversal"],
 	["apex undo --apply", "reverse the last setup without asking"],
 	["apex detect --json", "machine-readable inventory"]
 ];
 const environmentRows = [
-	["CALLSTACK_AUTH_TOKEN", "your callstack.ai key, needed by apex run and by every assistant you configure"],
+	...Object.values(PROVIDERS).map((provider) => [provider.envVar, `${provider.keyName}, needed by apex run and by every assistant you configure for ${provider.name}`]),
 	["Config directories", "XDG_CONFIG_HOME, CODEX_HOME, CLAUDE_CONFIG_DIR, PI_CODING_AGENT_DIR"],
 	["Undo journal", "APEX_STATE_DIR, XDG_STATE_HOME (default ~/.local/state/apex)"]
 ];
@@ -4574,7 +4734,7 @@ function helpText() {
 		""
 	];
 	return [
-		...wrap(`Apex CLI: point your coding assistants at ${MODEL}`, width$1).map((line) => bold(model(line))),
+		...wrap(`Apex CLI: point your coding assistants at Apex via ${Object.values(PROVIDERS).map((provider) => provider.name).join(" or the ")}`, width$1).map((line) => bold(model(line))),
 		"",
 		...section("Usage", usageRows.map(([lead, text]) => [
 			lead,
@@ -4614,9 +4774,11 @@ function readFlags(rest, command) {
 	const selected = values.assistants === void 0 ? null : [...new Set(values.assistants.split(",").map((id) => id.trim()))];
 	if (selected?.includes("")) throw new Error("--assistants needs at least one id.");
 	if (selected?.some((id) => !IDS.includes(id))) throw new Error(`Unknown assistant in --assistants. Choose from: ${IDS.join(", ")}`);
+	if (values.provider !== void 0 && !PROVIDER_IDS.includes(values.provider)) throw new Error(`Unknown --provider. Choose from: ${PROVIDER_IDS.join(", ")}`);
 	const json = Boolean(values.json);
 	return {
 		selected,
+		provider: values.provider === void 0 ? null : PROVIDERS[values.provider],
 		showDiff: !values["no-diff"],
 		json,
 		help: Boolean(values.help),
@@ -4627,7 +4789,7 @@ function readFlags(rest, command) {
 		})
 	};
 }
-async function planAll(assistants, options) {
+async function planAll(assistants, options, provider) {
 	return Promise.all(assistants.map(async (assistant) => {
 		const base = {
 			id: assistant.id,
@@ -4640,7 +4802,7 @@ async function planAll(assistants, options) {
 			steps: []
 		};
 		try {
-			base.files = (await planAssistant(assistant)).map((change) => describeFile(change, options));
+			base.files = (await planAssistant(assistant, provider)).map((change) => describeFile(change, options));
 			base.steps = base.files.flatMap((file) => file.change.steps ?? []);
 		} catch (error) {
 			base.error = error.message;
@@ -4648,7 +4810,7 @@ async function planAll(assistants, options) {
 		return base;
 	}));
 }
-const token = () => process.env.CALLSTACK_AUTH_TOKEN?.trim() || void 0;
+const token = (provider) => process.env[provider.envVar]?.trim() || void 0;
 const stamp = (iso) => iso.replace("T", " ").replace(/\.\d{3}Z$/, " UTC");
 const pendingOf = (entry) => entry.files.filter((file) => file.status !== "unchanged");
 async function applyAll(items, apply) {
@@ -4681,6 +4843,11 @@ const jsonAssistant = (entry, secret) => ({
 	...entry.steps.length ? { steps: entry.steps } : {},
 	files: entry.files.map((file) => jsonFile(file, secret))
 });
+const providerOptions = () => Object.values(PROVIDERS).map((provider) => ({
+	value: provider.id,
+	label: provider.name,
+	hint: `${provider.model}, key in ${provider.envVar}${token(provider) ? " (set)" : ""}`
+}));
 function pickerOptions(entries) {
 	return entries.map((entry) => {
 		const pending = pendingOf(entry).length;
@@ -4697,24 +4864,29 @@ async function commandInit(rest) {
 		printHelp();
 		return;
 	}
-	const secret = token();
+	const journalFile = journalPath();
+	if (flags.mode !== "preview") await readJournal(journalFile);
+	const batch = randomUUID();
+	const asking = !flags.json && flags.mode === "prompt";
+	let provider = flags.provider ?? defaultProvider();
+	if (!flags.json) intro("Apex CLI · Configure Apex for your favorite harness", MODE_NOTE[flags.mode]);
+	if (asking && !flags.provider) {
+		const picked = await selectProvider(providerOptions(), provider.id);
+		if (picked === false) return;
+		provider = PROVIDERS[picked];
+	}
+	const secret = token(provider);
 	const options = {
 		secret,
 		showDiff: flags.showDiff && !flags.json
 	};
-	const journalFile = journalPath();
-	if (flags.mode !== "preview") await readJournal(journalFile);
-	const batch = randomUUID();
-	const entries = await planAll(await detect(), options);
+	const entries = await planAll(await detect(), options, provider);
 	let ids = flags.selected ?? entries.filter((entry) => entry.detected).map((entry) => entry.id);
-	if (!flags.json) {
-		intro("Apex CLI · Configure callstack/Apex for your favorite harness", MODE_NOTE[flags.mode]);
-		if (flags.mode === "prompt" && !flags.selected) {
-			const defaults = entries.filter((entry) => entry.detected && !entry.manual && !entry.error && pendingOf(entry).length).map((entry) => entry.id);
-			const picked = await selectAssistants(pickerOptions(entries), defaults);
-			if (picked === false) return;
-			ids = picked;
-		}
+	if (asking && !flags.selected) {
+		const defaults = entries.filter((entry) => entry.detected && !entry.manual && !entry.error && pendingOf(entry).length).map((entry) => entry.id);
+		const picked = await selectAssistants(pickerOptions(entries), defaults, provider.model);
+		if (picked === false) return;
+		ids = picked;
 	}
 	const chosen = ids.map((id) => entries.find((entry) => entry.id === id));
 	const blocked = chosen.filter((entry) => entry.error);
@@ -4735,8 +4907,9 @@ async function commandInit(rest) {
 		console.log(JSON.stringify({
 			command: "init",
 			mode: flags.mode,
-			model: MODEL,
-			endpoint: BASE_URL,
+			provider: provider.id,
+			model: provider.model,
+			endpoint: provider.baseUrl,
 			applied: applied.length,
 			appliedPaths: applied,
 			...failure ? { error: failure.message } : {},
@@ -4748,7 +4921,7 @@ async function commandInit(rest) {
 			assistants: chosen.map((entry) => jsonAssistant(entry, secret)),
 			journal: journalFile,
 			undo: `${APEX} undo`,
-			envVar: "CALLSTACK_AUTH_TOKEN",
+			envVar: provider.envVar,
 			nextSteps: active.filter((entry) => RUNNABLE.includes(entry.id)).map((entry) => `${APEX} run ${entry.id}`)
 		}, null, 2));
 		if (blocked.length || failure) process.exitCode = 1;
@@ -4758,7 +4931,10 @@ async function commandInit(rest) {
 		outro(`No assistant selected, so nothing was written. Pick one with --assistants or run ${APEX} init interactively.`);
 		return;
 	}
-	if (flags.mode !== "prompt") prose(`${gray("setup for:")} ${chosen.map((entry) => entry.name + (entry.detected ? "" : " (not detected)")).join(", ")}`);
+	if (flags.mode !== "prompt") {
+		prose(`${gray("gateway:")} ${provider.name}`);
+		prose(`${gray("setup for:")} ${chosen.map((entry) => entry.name + (entry.detected ? "" : " (not detected)")).join(", ")}`);
+	}
 	if (blocked.length) {
 		plain("");
 		for (const entry of blocked) error(`${entry.name}: ${shortPath(entry.error)}`);
@@ -4786,23 +4962,24 @@ async function commandInit(rest) {
 	}
 	const apex = APEX !== "apex" && flags.mode === "prompt" && await installApex() ? "apex" : APEX;
 	for (const entry of active.filter((entry) => entry.steps.length)) section(`Finish setting up ${entry.name}`, entry.steps);
-	for (const entry of manual) section(`Manual setup for ${entry.name}: Apex CLI writes nothing`, MANUAL[entry.id]);
+	for (const entry of manual) section(`Manual setup for ${entry.name}: Apex CLI writes nothing`, manualSteps(entry.id, provider));
 	section("Environment", [
-		...secret ? [] : ["export CALLSTACK_AUTH_TOKEN=<your callstack.ai key>"],
-		secret ? `${green("✓")} CALLSTACK_AUTH_TOKEN is set in this shell` : `${yellow("!")}  CALLSTACK_AUTH_TOKEN is not set here yet`,
+		...secret ? [] : [`export ${provider.envVar}=<${provider.keyName.replace(/^your /, "")}>`],
+		secret ? `${green("✓")} ${provider.envVar} is set in this shell` : `${yellow("!")}  ${provider.envVar} is not set here yet`,
+		...!secret && provider.keysUrl ? [`Create a key at ${link(provider.keysUrl)}`] : [],
 		dim("Apex CLI never reads, stores or logs the key itself.")
 	]);
 	const runnable = active.map((entry) => entry.id).filter((id) => RUNNABLE.includes(id));
 	const rows = runnable.map((id) => ({
 		command: `${apex} run ${id}`,
-		expansion: runExpansion(id)
+		expansion: runExpansion(id, provider)
 	}));
 	const commandWidth = columnWidth(rows.map((row) => row.command));
-	section(runnable.length ? `Use these commands to run ${MODEL} with your selected harnesses:` : `Use ${MODEL} with your assistant:`, [
+	section(runnable.length ? `Use these commands to run ${provider.model} with your selected harnesses:` : `Use ${provider.model} with your assistant:`, [
 		...rows.flatMap((row) => wrap(row.expansion, width() - 3 - commandWidth).map((part, index) => `${pad(index ? "" : row.command, commandWidth)}${dim(part)}`)),
 		"",
-		`...or pick "${MODEL}" from the UI when setting up manually.`,
-		`For more instructions, visit: ${link(GUIDE_URL)}`,
+		`...or pick "${provider.model}" from the UI when setting up manually.`,
+		`For more instructions, visit: ${link(provider.guideUrl)}`,
 		...applied ? ["", `If you want to undo the changes, run ${bold(`${apex} undo`)}`] : [],
 		...apex === "apex" ? [] : ["", dim("For the short apex command: npm install -g @callstack/apex")]
 	]);
@@ -4813,16 +4990,17 @@ async function commandDetect(rest) {
 		printHelp();
 		return;
 	}
-	const secret = token();
+	const provider = flags.provider ?? defaultProvider();
+	const secret = token(provider);
 	const entries = await planAll(await detect(), {
 		secret,
 		showDiff: false
-	});
+	}, provider);
 	if (flags.json) {
 		console.log(JSON.stringify(entries.map((entry) => jsonAssistant(entry, secret)), null, 2));
 		return;
 	}
-	intro("Apex CLI · What this machine has", "Read only: nothing is created, edited or removed.");
+	intro(`Apex CLI · What this machine has for ${provider.name}`, "Read only: nothing is created, edited or removed.");
 	const rows = entries.map((entry) => {
 		const pending = pendingOf(entry).length;
 		return {
@@ -4848,7 +5026,7 @@ async function commandUndo(rest) {
 		printHelp();
 		return;
 	}
-	const secret = token();
+	const secret = token(defaultProvider());
 	const options = {
 		secret,
 		showDiff: flags.showDiff && !flags.json
@@ -4946,10 +5124,33 @@ async function commandUndo(rest) {
 		}
 	})).applied) outro("Undone. Configs are back to how they were. Restart any open assistant to reload them.");
 }
+function takeProvider(args) {
+	const [first, second] = args;
+	if (first === "--provider") return {
+		value: second ?? "",
+		rest: args.slice(2)
+	};
+	if (first?.startsWith("--provider=")) return {
+		value: first.slice(11),
+		rest: args.slice(1)
+	};
+	return {
+		value: null,
+		rest: args
+	};
+}
 async function commandRun(rest) {
-	const [id, ...passed] = rest;
+	const before = takeProvider(rest);
+	const [id, ...after] = before.rest;
 	if (!RUNNABLE.includes(id)) throw new Error(`apex run needs one of: ${RUNNABLE.join(", ")}. See ${APEX} --help.`);
-	const options = launchOptions(id);
+	const behind = before.value === null ? takeProvider(after) : {
+		value: null,
+		rest: after
+	};
+	const chosen = before.value ?? behind.value;
+	if (chosen !== null && !PROVIDER_IDS.includes(chosen)) throw new Error(`Unknown --provider. Choose from: ${PROVIDER_IDS.join(", ")}`);
+	const passed = behind.rest;
+	const options = launchOptions(id, process.env, chosen === null ? defaultProvider() : PROVIDERS[chosen]);
 	const binary = await executable(id);
 	if (!binary) throw new Error(`${id} was not found on PATH. Install it first, then run ${APEX} init.`);
 	if (process.platform === "win32" && /\.(cmd|bat)$/i.test(binary)) throw new Error("Windows .cmd/.bat launchers are not executed through a shell. Use WSL or launch the configured assistant directly.");
@@ -4972,7 +5173,8 @@ const completionsFor = (command) => COMMANDS[command].flags.map((flag) => ({
 	long: `--${flag}`,
 	short: FLAGS[flag].short ? `-${FLAGS[flag].short}` : null,
 	hint: FLAGS[flag].hint,
-	...command === "init" && flag === "assistants" ? { values: IDS } : {}
+	...command === "init" && flag === "assistants" ? { values: IDS } : {},
+	...flag === "provider" ? { values: PROVIDER_IDS } : {}
 }));
 const spellings = (entry) => [entry.long, entry.short].filter(Boolean);
 function completionScript(shell) {
@@ -4980,15 +5182,18 @@ function completionScript(shell) {
 	const words = (list) => list.join(" ");
 	if (shell === "zsh") return `#compdef apex
 _apex() {
-  local -a commands assistants runnable${taking.length ? " " : ""}${words(taking.map((name) => `${name}_flags`))}
+  local -a commands assistants providers runnable${taking.length ? " " : ""}${words(taking.map((name) => `${name}_flags`))}
   commands=(${words([...Object.entries(COMMANDS).map(([name, command]) => `'${name}:${command.summary}'`), ...GLOBALS.map(([flag, text]) => `'${flag}:${text}'`)])})
   assistants=(${words(IDS)})
+  providers=(${words(PROVIDER_IDS)})
   runnable=(${words(RUNNABLE)})
 ${taking.map((name) => `  ${name}_flags=(${words(completionsFor(name).flatMap((entry) => spellings(entry).map((word) => `'${word}:${entry.hint}'`)))})`).join("\n")}
   if (( CURRENT == 2 )); then
     _describe -t commands 'apex command' commands
   elif [[ "\${words[CURRENT - 1]}" == --assistants ]]; then
     _describe -t assistants 'assistant' assistants
+  elif [[ "\${words[CURRENT - 1]}" == --provider ]]; then
+    _describe -t providers 'gateway' providers
   else
     case "\${words[2]}" in
 ${taking.map((name) => `      ${name}) _describe -t flags 'option' ${name}_flags ;;`).join("\n")}
@@ -5003,6 +5208,7 @@ if [[ "\${funcstack[1]}" == _apex ]]; then _apex "$@"; else compdef _apex apex;
 	if (shell === "bash") return `_apex_completions() {
   local sub="\${COMP_WORDS[1]}" prev="\${COMP_WORDS[COMP_CWORD - 1]}" current="\${COMP_WORDS[COMP_CWORD]}"
   if [[ "$prev" == --assistants ]]; then COMPREPLY=( $(compgen -W "${words(IDS)}" -- "$current") ); return; fi
+  if [[ "$prev" == --provider ]]; then COMPREPLY=( $(compgen -W "${words(PROVIDER_IDS)}" -- "$current") ); return; fi
   if (( COMP_CWORD == 1 )); then
     COMPREPLY=( $(compgen -W "${words([...Object.keys(COMMANDS), ...GLOBALS.map(([flag]) => flag)])}" -- "$current") )
     return
@@ -5021,6 +5227,12 @@ complete -F _apex_completions apex
 		...GLOBALS.map(([flag, hint]) => `complete -c apex -f -n "__fish_use_subcommand" -l ${flag.slice(2)} -d '${hint}'`),
 		...taking.flatMap((name) => completionsFor(name).map((entry) => fish(`__fish_seen_subcommand_from ${name}`, entry))),
 		...RUNNABLE.map((id) => `complete -c apex -f -n "__fish_seen_subcommand_from run" -a ${id} -d 'assistant'`),
+		fish("__fish_seen_subcommand_from run", {
+			long: "--provider",
+			short: null,
+			hint: FLAGS.provider.hint,
+			values: PROVIDER_IDS
+		}),
 		...SHELLS.map((name) => `complete -c apex -f -n "__fish_seen_subcommand_from completion" -a ${name} -d 'shell'`)
 	].join("\n") + "\n";
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@callstack/apex",
-  "version": "0.4.1",
+  "version": "0.4.3",
   "description": "Detects the AI coding assistants installed on your machine and configures them to use Callstack Apex.",
   "license": "MIT",
   "keywords": [