@humain/terminal 0.1.0 → 0.1.1

package/docs/humain-terminal.md

@@ -55,6 +55,12 @@ Web settings and cached search content live under the HUMAIN agent directory. Th
 
 Maintainers can run `npm run check:web-access-upstream` from the repository root to detect a newer upstream release requiring source and dependency review.
 
+## HUMAIN Node
+
+Use `/login humain-node` to enter a masked API key and optional base URL (default `https://api.node.humain.com/v1`). Alternatively, set `HUMAIN_NODE_API_KEY` and optionally `HUMAIN_NODE_BASE_URL` in your environment.
+
+HUMAIN Terminal discovers the models available to that key before offering them in `/model`. Supported models use Node's Responses, Chat Completions, or Messages endpoint with normal streaming, tools, and reported token usage. Account spend, remaining quota, and pricing are not available from the supplied inference API. See [HUMAIN Node setup](providers.md#humain-node).
+
 ## Local Models
 
 Set `HUMAIN_TERMINAL_LOCAL_BASE_URL` to an OpenAI-compatible `/v1` endpoint. Models come from `HUMAIN_TERMINAL_LOCAL_MODELS` when set, otherwise HUMAIN Terminal performs a bounded `/v1/models` discovery request.

package/docs/providers.md

@@ -6,6 +6,7 @@ HUMAIN Terminal supports subscription-based providers via OAuth and API key prov
 
 - [Subscriptions](#subscriptions)
 - [API Keys](#api-keys)
+- [HUMAIN Node](#humain-node)
 - [Auth File](#auth-file)
 - [Cloud Providers](#cloud-providers)
 - [llama.cpp](#llamacpp)
@@ -72,6 +73,7 @@ humain-terminal
 | Ant Ling | `ANT_LING_API_KEY` | `ant-ling` |
 | Azure OpenAI Responses | `AZURE_OPENAI_API_KEY` | `azure-openai-responses` |
 | OpenAI | `OPENAI_API_KEY` | `openai` |
+| HUMAIN Node | `HUMAIN_NODE_API_KEY` | `humain-node` |
 | DeepSeek | `DEEPSEEK_API_KEY` | `deepseek` |
 | NVIDIA NIM | `NVIDIA_API_KEY` | `nvidia` |
 | Google Gemini | `GEMINI_API_KEY` | `google` |
@@ -104,7 +106,7 @@ humain-terminal
 | Xiaomi MiMo Token Plan (Amsterdam) | `XIAOMI_TOKEN_PLAN_AMS_API_KEY` | `xiaomi-token-plan-ams` |
 | Xiaomi MiMo Token Plan (Singapore) | `XIAOMI_TOKEN_PLAN_SGP_API_KEY` | `xiaomi-token-plan-sgp` |
 
-Reference for environment variables and `auth.json` keys: [`const envMap`](../../ai/src/env-api-keys.ts).
+Reference for built-in environment variables and `auth.json` keys: [`const envMap`](../../ai/src/env-api-keys.ts). HUMAIN Node is registered by the bundled HUMAIN extension.
 
 #### Auth File
 
@@ -184,6 +186,22 @@ The `key` field supports command execution, environment interpolation, and liter
 
 OAuth credentials are also stored here after `/login` and managed automatically.
 
+## HUMAIN Node
+
+Run `/login humain-node` (or select **HUMAIN Node** in `/login`). Enter the API key in the masked terminal prompt, then accept `https://api.node.humain.com/v1` or enter a custom base URL. Login validates the key with `GET /models` before saving credentials.
+
+To obtain a key, open your Team's **Users** page in HUMAIN Node, select the user for this workload and environment, and create an API key. Save it in your local environment as `HUMAIN_NODE_API_KEY` or in a secret manager. Never paste it into chat. If the environment variable is set, leave the login key prompt blank to use it without copying the key into `auth.json`.
+
+`HUMAIN_NODE_BASE_URL` optionally changes the default endpoint. A URL saved during login takes priority. Custom URLs must use HTTPS; HTTP is allowed for localhost testing.
+
+After login, use `/model` to select a discovered Node model and submit a prompt normally. For headless use, set `HUMAIN_NODE_API_KEY`, run `humain-terminal --list-models`, then pass a returned model ID with `humain-terminal --provider humain-node --model MODEL_ID -p "Say hello"`.
+
+Only models advertising a supported text API, streaming, and tool calling are offered for coding. Node's model metadata chooses Responses, Chat Completions, or Messages and supplies context/output limits and image-input support. Catalogs are refreshed from Node and kept in memory, scoped to the active key and URL; Node models are unavailable offline after a restart. Image generation and realtime sessions are outside the coding model picker.
+
+Token and cache usage reported by Node flows into the normal session statistics. The supplied API does not expose account spend, remaining quota, or model prices: zero-valued internal cost fields mean pricing is unavailable, not free usage. Use HUMAIN Node for billing and account usage reports.
+
+Failed requests include the supplied `X-Request-ID`. Temporary failures use bounded retries that respect `Retry-After`; quota and weekly-spend limits require action in Node. If a model becomes unavailable, open `/model` to refresh the catalog and select an available model. Run `/logout` and select **HUMAIN Node** to remove saved credentials; unset `HUMAIN_NODE_API_KEY` as well to disconnect an environment-based key.
+
 ## Cloud Providers
 
 ### Azure OpenAI

package/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
 	"name": "@humain/terminal",
-	"version": "0.1.0",
+	"version": "0.1.1",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "@humain/terminal",
-			"version": "0.1.0",
+			"version": "0.1.1",
 			"license": "MIT",
 			"dependencies": {
 				"@humain/terminal-contracts": "0.1.0",

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@humain/terminal",
-	"version": "0.1.0",
+	"version": "0.1.1",
 	"description": "Coding agent CLI with read, bash, edit, write tools and session management",
 	"type": "module",
 	"humainTerminalConfig": {