From ccb68cf0f5fa7297b2d7eb7e49f6c1f7dead96f2 Mon Sep 17 00:00:00 2001
From: Adeel Ahmad
Date: Sun, 18 Oct 2026 12:06:22 +0000
Subject: [PATCH 1/2] feat(scaling): add providers and make groups in Settings
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Settings › Auto scaling lists the providers with their last check in
words, Check again, and Remove with Undo (the delete waits a few seconds,
or runs as the page is left, since sealed credentials can't be put back).
Add a provider picks a driver and fills a form drawn from its fields:
defaults filled, the name focused, secrets as password fields sent in
`secrets`, lists one per line or ticked from their options, Enter saves,
and the driver's extra and licence named; a failed check is shown in
words. New group on a provider asks for a name, min, max, a budget only
when the driver is priced, and the driver's template; rules stay with
the assistant, MCP and the API for now.
Co-Authored-By: Claude Opus 5.5 (1M context)
Claude-Session: https://claude.ai/code/session_01ALjNbLobvPe7VwvDL8j7V3
---
README.md | 17 +-
changelog.d/scaling-forms.md | 6 +
docs/requirements.md | 1 +
docs/scaling.md | 36 +-
.../__tests__/settings-scaling.test.tsx | 174 +++++-
.../components/settings/bodies.tsx | 8 +-
.../components/settings/scaling-panel.tsx | 586 +++++++++++++++++-
7 files changed, 801 insertions(+), 27 deletions(-)
create mode 100644 changelog.d/scaling-forms.md
diff --git a/README.md b/README.md
index 80b144ff2..7ccf31922 100644
--- a/README.md
+++ b/README.md
@@ -237,13 +237,12 @@ listed under [In progress](#in-progress), and what comes next under [Planned](#p
[the meter](docs/budgets.md#the-meter-and-the-costs-page); [checks](docs/requirements.md#see-what-it-costs))
- **Scale out.** Scaling groups add workers when work waits and let them go when it doesn't, by target, step and
schedule rules, within min and max and a monthly budget, with every launch, join and hour of cost in the activity
- ledger. A group launches nothing until you approve its terms in Settings › Auto scaling; agents propose and manage
- groups over the API, MCP and the assistant, and never approve one. It runs more workers on this machine today;
- AWS EC2 (spot or on demand), DigitalOcean, Hetzner Cloud, Proxmox VE and Linux hosts over SSH providers are in
- progress (built, with machines that join over the mesh to transcribe and tell voices apart, not yet tried on a real
- cloud), so is RunPod for GPUs that serve Lens's model calls while they run, and more model hosts are planned.
- ([Auto scaling](docs/scaling.md);
- [checks](docs/requirements.md#scale-out))
+ ledger. You add providers and make groups in Settings › Auto scaling, and a group launches nothing until you approve
+ its terms there; agents propose and manage groups over the API, MCP and the assistant, and never approve one. It
+ runs more workers on this machine today; AWS EC2 (spot or on demand), DigitalOcean, Hetzner Cloud, Proxmox VE and
+ Linux hosts over SSH providers are in progress (built, with machines that join over the mesh to transcribe and tell
+ voices apart, not yet tried on a real cloud), so is RunPod for GPUs that serve Lens's model calls while they run,
+ and more model hosts are planned. ([Auto scaling](docs/scaling.md); [checks](docs/requirements.md#scale-out))
- **Run on more machines.** Add a Mac with a GPU, a box for the assistant's code or a server with more disk, and
Lens joins it over a private network of its own, a Nebula mesh it issues the certificates for: one install line
with a one-time token, and the machine needs nothing else. Worker nodes run jobs as their own user of the
@@ -320,8 +319,8 @@ Next, and not in a pull request yet; each links to where it is tracked:
topics, recordings and the other things that have pages ([notes](docs/notes.md)).
- **Local models:** GPU builds of llama.cpp fetched by Lens, and local embedding models
([local models](docs/local-models.md)).
-- **Auto scaling:** spreading model calls across GPUs, more model hosts (Lambda, Hugging Face, SageMaker, Bedrock),
- and making providers and groups in Settings ([auto scaling](docs/scaling.md)).
+- **Auto scaling:** spreading model calls across GPUs, and more model hosts (Lambda, Hugging Face, SageMaker,
+ Bedrock) ([auto scaling](docs/scaling.md)).
- **Speech providers:** their summaries, chapters and entities as optional imports, and per-minute prices in the
activity ledger ([speech providers](docs/speech-providers.md)).
- **Storage:** uploads and assets' media in the file store, and a storage key to download
diff --git a/changelog.d/scaling-forms.md b/changelog.d/scaling-forms.md
new file mode 100644
index 000000000..b4a0cc37d
--- /dev/null
+++ b/changelog.d/scaling-forms.md
@@ -0,0 +1,6 @@
+- **Add providers and make scaling groups in Settings.** Settings › Auto scaling lists the providers with their last
+ check in words, Check again, and Remove with Undo. Add a provider picks the kind, then fills a form drawn from the
+ driver's fields (defaults filled, credentials as password fields, Enter saves) and shows the check it runs. New group
+ on a provider asks for a name, the minimum and maximum, a monthly budget when the provider charges, and what each
+ unit is; a group made there is approved as it's made. Rules stay with the assistant, MCP and the API for now
+ ([In Settings](docs/scaling.md#in-settings)).
diff --git a/docs/requirements.md b/docs/requirements.md
index 03eefa0ad..8c51436e6 100644
--- a/docs/requirements.md
+++ b/docs/requirements.md
@@ -287,6 +287,7 @@ request, to another check that proves the same thing; loosening a requirement or
| The scaler launches units, sees them join, drains busy workers before letting them go, and fails and pauses on units that don't come up | [test_scaling.py](https://github.com/adeelahmad/lense/blob/main/fastapi_backend/tests/api/test_scaling.py) `test_the_scaler_launches_joins_drains_and_terminates`, `test_a_unit_that_never_joins_fails_and_failures_pause_the_group`, `test_protected_and_terminated_units` |
| Each unit's cost goes into the activity ledger as an estimate, and a group stops at its budget's cap | [test_scaling.py](https://github.com/adeelahmad/lense/blob/main/fastapi_backend/tests/api/test_scaling.py) `test_the_scaler_launches_joins_drains_and_terminates`, `test_at_its_budgets_cap_a_group_stops` |
| Settings › Auto scaling shows each group's terms in words and approves them in one click | [settings-scaling.test.tsx](https://github.com/adeelahmad/lense/blob/main/nextjs-frontend/__tests__/settings-scaling.test.tsx) `shows each group's terms and approves them in one click`, `puts the terms and the state in words` |
+| Settings › Auto scaling adds a provider from its driver's fields (credentials sent apart, never shown), shows its check in words, removes it with Undo, and makes a group, with a budget when the provider charges | [settings-scaling.test.tsx](https://github.com/adeelahmad/lense/blob/main/nextjs-frontend/__tests__/settings-scaling.test.tsx) `adds a provider from its driver's fields, with the secret sent apart`, `shows a failed check in words, after saving and in the list`, `makes a group with a budget on a priced provider`, `asks no budget on a free provider and ticks a list's options`, `removes a provider with Undo instead of a question` |
| This machine runs more workers at no cost | [test_scaling.py](https://github.com/adeelahmad/lense/blob/main/fastapi_backend/tests/api/test_scaling.py) `test_this_machine_runs_more_workers_at_no_cost` |
| DigitalOcean and Hetzner Cloud: credentials checked, list prices, machines launched with a one-use join token, watched, found by tag and deleted; a group waits for the mesh's install line before launching | [test_scaling_clouds.py](https://github.com/adeelahmad/lense/blob/main/fastapi_backend/tests/api/test_scaling_clouds.py) `test_a_digitalocean_group_waits_for_the_install_line_then_launches_droplets`, `test_hetzner_prices_in_euros_and_its_errors_read_as_hetzners`, `test_the_cloud_drivers_are_listed_and_need_the_mesh` |
| A group of machines elsewhere counts only the jobs a worker node takes, and never launches one for others | [test_scaling_clouds.py](https://github.com/adeelahmad/lense/blob/main/fastapi_backend/tests/api/test_scaling_clouds.py) `test_a_group_elsewhere_counts_only_the_jobs_a_node_takes` |
diff --git a/docs/scaling.md b/docs/scaling.md
index da6d8c7a7..8284984e8 100644
--- a/docs/scaling.md
+++ b/docs/scaling.md
@@ -7,14 +7,14 @@ launched. Everything it does is in the [activity ledger](activity.md), and its c
[budget](budgets.md).
Status: in progress. Built: groups, rules, approvals, budgets, the scaler, the API, MCP and assistant tools,
-Settings › Auto scaling (the switch, and each group's terms to approve or pause), and **This machine**, a provider that
-runs more `lens worker` processes where Lens runs. In progress: **AWS EC2** (spot or on demand), **DigitalOcean**
-and **Hetzner Cloud** providers are built (credentials checked, prices, launching, watching and deleting machines),
-and their machines join over the [mesh](mesh.md) as worker nodes that transcribe and tell voices apart; so are
-**Proxmox VE**, clones of a VM template on your own server, and **Linux hosts over SSH**, workers started on machines
-you have. They haven't been tried on a real cloud machine yet. Also in progress: **RunPod**, GPUs that serve a model
-with vLLM, which Lens's model calls go to while one runs. Planned: more model hosts (Lambda, Hugging Face, SageMaker,
-Bedrock); and making providers and groups in Settings.
+Settings › Auto scaling (the switch, [adding providers and making groups](#in-settings), and each group's terms to
+approve or pause), and **This machine**, a provider that runs more `lens worker` processes where Lens runs. In
+progress: **AWS EC2** (spot or on demand), **DigitalOcean** and **Hetzner Cloud** providers are built (credentials
+checked, prices, launching, watching and deleting machines), and their machines join over the [mesh](mesh.md) as
+worker nodes that transcribe and tell voices apart; so are **Proxmox VE**, clones of a VM template on your own server,
+and **Linux hosts over SSH**, workers started on machines you have. They haven't been tried on a real cloud machine
+yet. Also in progress: **RunPod**, GPUs that serve a model with vLLM, which Lens's model calls go to while one runs.
+Planned: more model hosts (Lambda, Hugging Face, SageMaker, Bedrock).
Auto scaling is **off** until an admin turns it on, and a group launches nothing until a person approves its terms
signed in to the web app. Agents can propose and manage groups, but never approve one.
@@ -40,6 +40,23 @@ One process runs the scaler at a time: the API and every `lens worker` (except t
`--no-schedule`, as scaled workers are) take turns holding its lease. While scaling is off and nothing is left to let
go, the scaler doesn't even take the lease.
+## In Settings
+
+**Settings › Auto scaling** lists the providers, each with its kind, which credentials are set and whether its last
+check passed (and why not, in words when it failed), with **Check again** and **Remove**. Remove takes it out of the
+list at once with **Undo** on the notice; it's deleted a few seconds later, or as you leave the page, since its sealed
+credentials can't be put back once it's gone. A provider a group uses can't be removed until the group goes.
+
+**Add a provider** asks for the kind (each with what it does), then a form drawn from the driver's `fields`: defaults
+filled in, the name first and focused, credentials as password fields (sent as `secrets`, never shown again), lists
+one per line, on and off settings as checkboxes, and Enter saves. A driver that needs packages says which, and their
+licence, before it fetches them on the first save. The check runs as it's saved; when it fails, the dialog says why.
+
+**New group** on a provider asks for a name (` workers`), the minimum (0) and maximum (1), a monthly budget
+in USD only when the provider charges money (then it's required), and the driver's `template` with its defaults. A
+group made here is approved as it's made. Its rules are set by the assistant, MCP or the API for now; without rules it
+keeps its minimum running.
+
## Providers
```http
@@ -271,4 +288,5 @@ preemptible and spot instances at clouds other than AWS and RunPod; draining an
(the node agent reading the instance's metadata) and handing its job back at once; following the spot price as it
changes; AWS sign-in with an instance role or IAM Identity Center instead of keys; mixed instance types per group;
warm pools; predictive scaling; health checks beyond
-joining; prices fetched from providers' pricing APIs.
+joining; prices fetched from providers' pricing APIs. In Settings: rules, changing a provider or a group, removing a
+group, and a provider's price and note.
diff --git a/nextjs-frontend/__tests__/settings-scaling.test.tsx b/nextjs-frontend/__tests__/settings-scaling.test.tsx
index 87dcd1c1b..fd373ae51 100644
--- a/nextjs-frontend/__tests__/settings-scaling.test.tsx
+++ b/nextjs-frontend/__tests__/settings-scaling.test.tsx
@@ -1,9 +1,9 @@
import { QueryClient, QueryClientProvider } from "@tanstack/react-query";
-import { fireEvent, render, screen, waitFor } from "@testing-library/react";
+import { act, fireEvent, render, screen, waitFor } from "@testing-library/react";
import "@testing-library/jest-dom";
import { Scaling } from "@/app/openapi-client";
-import type { Group, Provider } from "@/app/openapi-client/types.gen";
+import type { DriverSpec, Group, Provider } from "@/app/openapi-client/types.gen";
import { groupState, termsText } from "@/components/settings/scaling-panel";
import { SettingsSection } from "@/components/settings/section";
import type { SettingsView } from "@/components/settings/model";
@@ -11,13 +11,24 @@ import { TooltipProvider } from "@/components/ui/tooltip";
jest.mock("@/app/openapi-client", () => ({
Admin: { updateSettings: jest.fn() },
- Scaling: { listGroups: jest.fn(), listProviders: jest.fn(), approveGroup: jest.fn(), pauseGroup: jest.fn() },
+ Scaling: {
+ listGroups: jest.fn(),
+ listProviders: jest.fn(),
+ listDrivers: jest.fn(),
+ addProvider: jest.fn(),
+ checkProvider: jest.fn(),
+ removeProvider: jest.fn(),
+ createGroup: jest.fn(),
+ approveGroup: jest.fn(),
+ pauseGroup: jest.fn(),
+ },
Metadata: {},
}));
jest.mock("next-auth/react", () => ({ useSession: () => ({ data: { accessToken: "t" } }) }));
jest.mock("next/navigation", () => ({ useRouter: () => ({ push: jest.fn() }) }));
jest.mock("@/lib/hooks/session", () => ({ useArchive: () => ({ namespaces: [] }) }));
-jest.mock("@/components/ui/toast", () => ({ useToast: () => jest.fn() }));
+const mockToast = jest.fn();
+jest.mock("@/components/ui/toast", () => ({ useToast: () => mockToast }));
const ok = (data: unknown) => Promise.resolve({ data, response: { ok: true, status: 200 } });
const VIEW = {
@@ -27,6 +38,34 @@ const PROVIDERS = [
{ id: 1, name: "Here", driver: "local", priced: false, settings: {}, secrets: {} },
{ id: 2, name: "Cloud", driver: "fake", priced: true, settings: {}, secrets: { api_key: true } },
] as Provider[];
+const DRIVERS = [
+ {
+ id: "local",
+ kind: "node",
+ label: "This machine",
+ about: "Runs more Lens workers here.",
+ priced: false,
+ ready: true,
+ fields: {},
+ template: { steps: { label: "Steps", type: "list", options: ["embed", "transcribe"] } },
+ },
+ {
+ id: "fake",
+ kind: "node",
+ label: "Fake cloud",
+ about: "Machines on a pretend cloud.",
+ priced: true,
+ extra: "fakecloud",
+ license: "MIT",
+ ready: false,
+ fields: {
+ region: { label: "Region", type: "string", required: true, default: "ams3" },
+ token: { label: "API token", type: "string", required: true, secret: true },
+ private: { label: "Private network", type: "boolean", default: true },
+ },
+ template: { size: { label: "Size", type: "string", default: "small" } },
+ },
+] as DriverSpec[];
const GROUP = {
id: 7,
name: "Transcribers",
@@ -61,7 +100,9 @@ function show() {
}
beforeEach(() => {
+ jest.clearAllMocks();
(Scaling.listProviders as jest.Mock).mockImplementation(() => ok(PROVIDERS));
+ (Scaling.listDrivers as jest.Mock).mockImplementation(() => ok(DRIVERS));
});
describe("auto scaling settings", () => {
@@ -93,4 +134,129 @@ describe("auto scaling settings", () => {
"2 of 3 running, 1 joining",
);
});
+
+ it("adds a provider from its driver's fields, with the secret sent apart", async () => {
+ (Scaling.listGroups as jest.Mock).mockImplementation(() => ok([]));
+ (Scaling.addProvider as jest.Mock).mockImplementation(({ body }) =>
+ ok({ id: 3, ...body, priced: true, secrets: { token: true }, check: { ok: true } }),
+ );
+ show();
+ fireEvent.click(await screen.findByRole("button", { name: "Add a provider" }));
+ fireEvent.click(await screen.findByRole("button", { name: /Fake cloud/ }));
+ expect(screen.getByText(/Needs the fakecloud packages \(MIT\)/)).toBeInTheDocument();
+ expect(screen.getByLabelText("Name")).toHaveValue("Fake cloud");
+ expect(screen.getByLabelText("Name")).toHaveFocus();
+ expect(screen.getByLabelText(/Region/)).toHaveValue("ams3");
+ const token = screen.getByLabelText(/API token/);
+ expect(token).toHaveAttribute("type", "password");
+ fireEvent.change(token, { target: { value: "s3cret" } });
+ fireEvent.keyDown(token, { key: "Enter" });
+ await waitFor(() => expect(Scaling.addProvider).toHaveBeenCalled());
+ expect((Scaling.addProvider as jest.Mock).mock.calls[0][0].body).toEqual({
+ name: "Fake cloud",
+ driver: "fake",
+ settings: { region: "ams3", private: true },
+ secrets: { token: "s3cret" },
+ });
+ await waitFor(() => expect(mockToast).toHaveBeenCalledWith(expect.objectContaining({ title: "Added Fake cloud" })));
+ });
+
+ it("shows a failed check in words, after saving and in the list", async () => {
+ (Scaling.listGroups as jest.Mock).mockImplementation(() => ok([]));
+ (Scaling.listProviders as jest.Mock).mockImplementation(() =>
+ ok([{ ...PROVIDERS[1], secrets: { token: true }, check: { ok: false, error: "the token was refused" } }]),
+ );
+ (Scaling.addProvider as jest.Mock).mockImplementation(({ body }) =>
+ ok({
+ id: 3,
+ ...body,
+ priced: true,
+ secrets: { token: true },
+ check: { ok: false, error: "the region is unknown" },
+ }),
+ );
+ (Scaling.checkProvider as jest.Mock).mockImplementation(() => ok({ ok: false, error: "the token was refused" }));
+ show();
+ expect(await screen.findByText("Check failed: The token was refused.")).toBeInTheDocument();
+ expect(screen.getByText(/API token set/)).toBeInTheDocument();
+ fireEvent.click(screen.getByRole("button", { name: "Check Cloud again" }));
+ await waitFor(() => expect(Scaling.checkProvider).toHaveBeenCalled());
+ expect((Scaling.checkProvider as jest.Mock).mock.calls[0][0].path).toEqual({ pid: 2 });
+
+ fireEvent.click(screen.getByRole("button", { name: "Add a provider" }));
+ fireEvent.click(await screen.findByRole("button", { name: /Fake cloud/ }));
+ fireEvent.click(screen.getByRole("button", { name: "Add provider" }));
+ expect(await screen.findByText("API token is needed.")).toBeInTheDocument();
+ expect(Scaling.addProvider).not.toHaveBeenCalled();
+ fireEvent.change(screen.getByLabelText(/API token/), { target: { value: "s3cret" } });
+ fireEvent.click(screen.getByRole("button", { name: "Add provider" }));
+ expect(await screen.findByText("It’s saved, but its check failed.")).toBeInTheDocument();
+ expect(screen.getByText(/The region is unknown\./)).toBeInTheDocument();
+ });
+
+ it("makes a group with a budget on a priced provider", async () => {
+ (Scaling.listGroups as jest.Mock).mockImplementation(() => ok([]));
+ (Scaling.createGroup as jest.Mock).mockImplementation(({ body }) =>
+ ok({ ...GROUP, ...body, budget: { ...GROUP.budget, usd: body.budget }, approved: true }),
+ );
+ show();
+ fireEvent.click(await screen.findByRole("button", { name: "New group on Cloud" }));
+ expect(screen.getByLabelText("Name")).toHaveValue("Cloud workers");
+ expect(screen.getByLabelText("Minimum")).toHaveValue(0);
+ expect(screen.getByLabelText("Maximum")).toHaveValue(1);
+ expect(screen.getByLabelText(/Size/)).toHaveValue("small");
+ expect(screen.getByText(/set by the assistant, MCP or the API for now/)).toBeInTheDocument();
+ fireEvent.click(screen.getByRole("button", { name: "Make group" }));
+ expect(await screen.findByText(/A budget is needed/)).toBeInTheDocument();
+ expect(Scaling.createGroup).not.toHaveBeenCalled();
+ fireEvent.change(screen.getByLabelText("Budget, USD a month"), { target: { value: "25" } });
+ fireEvent.change(screen.getByLabelText("Maximum"), { target: { value: "3" } });
+ fireEvent.click(screen.getByRole("button", { name: "Make group" }));
+ await waitFor(() => expect(Scaling.createGroup).toHaveBeenCalled());
+ expect((Scaling.createGroup as jest.Mock).mock.calls[0][0].body).toEqual({
+ name: "Cloud workers",
+ provider: 2,
+ min: 0,
+ max: 3,
+ template: { size: "small" },
+ budget: 25,
+ });
+ });
+
+ it("asks no budget on a free provider and ticks a list's options", async () => {
+ (Scaling.listGroups as jest.Mock).mockImplementation(() => ok([]));
+ (Scaling.createGroup as jest.Mock).mockImplementation(({ body }) => ok({ ...GROUP, ...body, budget: null }));
+ show();
+ fireEvent.click(await screen.findByRole("button", { name: "New group on Here" }));
+ expect(screen.queryByLabelText("Budget, USD a month")).not.toBeInTheDocument();
+ fireEvent.click(screen.getByRole("checkbox", { name: "transcribe" }));
+ fireEvent.keyDown(screen.getByLabelText("Name"), { key: "Enter" });
+ await waitFor(() => expect(Scaling.createGroup).toHaveBeenCalled());
+ expect((Scaling.createGroup as jest.Mock).mock.calls[0][0].body).toEqual({
+ name: "Here workers",
+ provider: 1,
+ min: 0,
+ max: 1,
+ template: { steps: ["transcribe"] },
+ });
+ });
+
+ it("removes a provider with Undo instead of a question", async () => {
+ (Scaling.listGroups as jest.Mock).mockImplementation(() => ok([GROUP]));
+ (Scaling.removeProvider as jest.Mock).mockImplementation(() => ok({ ok: true }));
+ const { unmount } = show();
+ fireEvent.click(await screen.findByRole("button", { name: "Remove Here" }));
+ expect(screen.queryByText("Here")).not.toBeInTheDocument();
+ const undo = mockToast.mock.calls[mockToast.mock.calls.length - 1][0];
+ expect(undo.title).toBe("Removed Here");
+ act(() => undo.action.onClick());
+ expect(screen.getByText("Here")).toBeInTheDocument();
+ expect(Scaling.removeProvider).not.toHaveBeenCalled();
+ // Cloud runs Transcribers, so it can't go yet
+ expect(screen.getByRole("button", { name: "Remove Cloud" })).toHaveAttribute("aria-disabled", "true");
+ // leaving the page removes what waits for Undo at once
+ fireEvent.click(screen.getByRole("button", { name: "Remove Here" }));
+ unmount();
+ expect((Scaling.removeProvider as jest.Mock).mock.calls[0][0].path).toEqual({ pid: 1 });
+ });
});
diff --git a/nextjs-frontend/components/settings/bodies.tsx b/nextjs-frontend/components/settings/bodies.tsx
index e66ea4e9e..c1c73d501 100644
--- a/nextjs-frontend/components/settings/bodies.tsx
+++ b/nextjs-frontend/components/settings/bodies.tsx
@@ -28,7 +28,7 @@ import {
type ModelProvider,
} from "@/components/settings/model-providers";
import { SandboxGatewaySwitch } from "@/components/settings/sandbox-gateway";
-import { ScalingGroups } from "@/components/settings/scaling-panel";
+import { ScalingGroups, ScalingProviders } from "@/components/settings/scaling-panel";
import { SignInProviders } from "@/components/settings/sign-in-providers";
import { UpdatesPanel } from "@/components/settings/updates-panel";
import { settingHref } from "@/components/settings/find";
@@ -620,6 +620,8 @@ export function SectionBody({ ctx }: { ctx: BodyCtx }) {
return (
<>
+ Providers
+ Groups
A group launches nothing until you approve its terms here; agents (the assistant, MCP and API keys) make and
manage groups but can’t approve them. Costs are estimates from list prices, counted against each group’s
- budget. Today groups run more workers on this machine; cloud providers are planned.
+ budget. Today groups run more workers on this machine, and RunPod GPUs serve Lens’s model calls; AWS,
+ DigitalOcean, Hetzner Cloud, Proxmox VE and SSH providers launch machines that join over the mesh as worker
+ nodes, once the mesh is on in Settings › Network.
>
);
diff --git a/nextjs-frontend/components/settings/scaling-panel.tsx b/nextjs-frontend/components/settings/scaling-panel.tsx
index f164a4fca..5a1608887 100644
--- a/nextjs-frontend/components/settings/scaling-panel.tsx
+++ b/nextjs-frontend/components/settings/scaling-panel.tsx
@@ -1,14 +1,18 @@
"use client";
import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
+import { useEffect, useRef, useState } from "react";
import { Scaling } from "@/app/openapi-client";
-import type { Group, Provider } from "@/app/openapi-client/types.gen";
+import type { DriverField, DriverSpec, Group, Provider } from "@/app/openapi-client/types.gen";
import { Banner } from "@/components/ui/banner";
import { Button } from "@/components/ui/button";
+import { Dialog } from "@/components/ui/dialog";
+import { Checkbox, Field, Input, Select, Textarea } from "@/components/ui/field";
import { SkeletonRows } from "@/components/ui/states";
import { useToast } from "@/components/ui/toast";
import { data, useApiClient } from "@/lib/api/browser";
+import { cn } from "@/lib/utils";
/** One of the API's reasons as a sentence: capitalised, with a full stop. */
function sentence(s: string): string {
@@ -76,8 +80,8 @@ export function ScalingGroups() {
if (!list.length)
return (
- No scaling groups yet. Ask the assistant to set one up, or make one over the API or MCP; it shows here for you
- to approve.
+ No scaling groups yet. Make one with New group on a provider above, or ask the assistant to set one up (or make
+ one over the API or MCP); it shows here for you to approve.
);
return (
@@ -115,3 +119,579 @@ export function ScalingGroups() {
);
}
+
+type Fields = Record;
+/** What a person has typed for a driver's fields: text, a checkbox, or the options ticked. */
+export type SpecForm = Record;
+
+const picks = (f: DriverField) => (f.options ?? []).map(String);
+const checkOf = (p: Provider) => p.check as { ok?: boolean; error?: string } | null | undefined;
+
+/** A driver's fields as a form, with their defaults filled in. */
+export function specForm(fields: Fields): SpecForm {
+ return Object.fromEntries(
+ Object.entries(fields).map(([k, f]) => {
+ const d = f.default;
+ if (f.type === "boolean") return [k, d === true];
+ if (f.type === "list")
+ return [k, f.options?.length ? (Array.isArray(d) ? d.map(String) : []) : Array.isArray(d) ? d.join("\n") : ""];
+ return [k, d == null ? "" : String(d)];
+ }),
+ );
+}
+
+/** The form as the API takes it: settings (or a template) and, apart, the secrets; or what's wrong, by field. A field
+ * left empty is left out, so the driver's default applies. */
+export function specValues(
+ fields: Fields,
+ form: SpecForm,
+): { values: Record; secrets: Record; errors: Record } {
+ const values: Record = {};
+ const secrets: Record = {};
+ const errors: Record = {};
+ for (const [k, f] of Object.entries(fields)) {
+ const raw = form[k];
+ let v: unknown;
+ if (f.type === "boolean") v = raw === true;
+ else if (f.type === "list") {
+ const xs = Array.isArray(raw)
+ ? raw
+ : String(raw ?? "")
+ .split(/[\n,]/)
+ .map((x) => x.trim())
+ .filter(Boolean);
+ v = xs.length ? xs : undefined;
+ } else {
+ const t = String(raw ?? "").trim();
+ if (t && (f.type === "integer" || f.type === "number")) {
+ const n = Number(t);
+ if (f.type === "integer" ? !Number.isInteger(n) : !Number.isFinite(n)) {
+ errors[k] = f.type === "integer" ? "A whole number." : "A number.";
+ continue;
+ }
+ v = n;
+ } else v = t || undefined;
+ }
+ if (v === undefined) {
+ if (f.required) errors[k] = `${f.label} is needed.`;
+ } else if (f.secret) secrets[k] = String(v);
+ else values[k] = v;
+ }
+ return { values, secrets, errors };
+}
+
+/** A driver's fields, drawn from its spec: secrets as passwords, lists one per line (or ticked from their options),
+ * booleans as checkboxes, a choice of options as a select. */
+export function SpecFields({
+ fields,
+ form,
+ onChange,
+ errors,
+}: {
+ fields: Fields;
+ form: SpecForm;
+ onChange: (f: SpecForm) => void;
+ errors: Record;
+}) {
+ const set = (k: string, v: string | boolean | string[]) => onChange({ ...form, [k]: v });
+ return (
+ <>
+ {Object.entries(fields).map(([k, f]) => {
+ const help = f.help ? sentence(f.help) : undefined;
+ if (f.type === "boolean")
+ return (
+