Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 900+ contributors
TypeScript a2aai-agentsai-gatewayanthropicclaudeclaude-codeclinecodexcopilotcursordeepseekfree-aigeminikimillm-gatewaymcpopenaiopenai-proxyqwentoken-saver
/**
* Pure merge of registry / synced / custom model rows for the provider detail
* dashboard (and thus Test All targets). Server-confirmed authoritative catalogs
* exclude retired static/imported rows while preserving operator-owned models.
*/
import { appendSyncedEffortVariants } from "@omniroute/open-sse/utils/syncedEffortVariants";
import { ensureCursorAutoCatalogEntry } from "@/lib/providerModels/cursorAutoCatalog";
import { mergeModelsWithCustomPrecedence } from "@/lib/providers/modelMetadataPrecedence";
import {
providerUsesCuratedModelsOnly,
providerUsesExclusiveSyncedListing,
} from "@/lib/providers/modelListingCapability";
export type ProviderListingModel = {
id: string;
name?: string;
source?: string;
[key: string]: unknown;
};
/** Apply the same membership rule to compatible/passthrough rows, including alias-only rows. */
export function filterUnavailableModelRows<T extends { modelId: string }>(
rows: T[],
syncedModels: Array<{ id?: string }>,
customModels: Array<{ id?: string; source?: string }>,
authoritative: boolean
): T[] {
if (!authoritative) return rows;
const allowed = new Set([
...syncedModels.map((model) => model.id),
...customModels.filter((model) => model.source !== "imported").map((model) => model.id),
]);
return rows.filter((row) => allowed.has(row.modelId));
}
export type MergeProviderModelListingInput = {
providerId: string;
registryModels: Array<{ id: string; name?: string }>;
syncedModels: Array<{ id: string; name?: string; [key: string]: unknown }>;
customModels: Array<{ id: string; name?: string; source?: string; [key: string]: unknown }>;
usesCuratedModelsOnly?: boolean;
syncedCatalogAuthoritative?: boolean;
};
function normalizeCustomSource(source: unknown): "imported" | "custom" {
return source === "imported" ? "imported" : "custom";
}
function dedupeById(models: ProviderListingModel[]): ProviderListingModel[] {
const deduped = new Map<string, ProviderListingModel>();
for (const m of models) {
if (m.id && !deduped.has(m.id)) deduped.set(m.id, m);
}
return Array.from(deduped.values());
}
export function mergeProviderModelListing(
input: MergeProviderModelListingInput
): ProviderListingModel[] {
const curated =
input.usesCuratedModelsOnly === true || providerUsesCuratedModelsOnly(input.providerId);
const synced = curated ? [] : input.syncedModels.filter((m) => m?.id);
const custom = curated ? [] : input.customModels.filter((m) => m?.id);
const exclusive =
!curated &&
(input.syncedCatalogAuthoritative ??
(providerUsesExclusiveSyncedListing(input.providerId) && synced.length > 0));
if (exclusive) {
const cursor = ["cursor", "cu"].includes(input.providerId.trim().toLowerCase());
const registryById = new Map(input.registryModels.map((model) => [model.id, model]));
const liveModels = synced.map((model) => ({
...(registryById.get(model.id) || {}),
...model,
id: model.id,
name: model.name || model.id,
owned_by: cursor ? "cursor" : input.providerId,
source: "imported",
}));
const withAuto =
cursor && liveModels.length > 0 ? ensureCursorAutoCatalogEntry(liveModels) : liveModels;
const liveIds = new Set(withAuto.map((model) => model.id));
const normalizedCustom = custom
.filter((model) => model.source !== "imported" || liveIds.has(model.id))
.map((model) => ({
...model,
id: model.id,
name: model.name || model.id,
source: normalizeCustomSource(model.source),
}));
const exclusiveMerged = dedupeById(mergeModelsWithCustomPrecedence(withAuto, normalizedCustom));
// Codex is exclusive since #15132, but the tier variants #13224 derives from the
// account's own discovered reasoning levels are live data, not static aliases:
// keep them, as /v1/models does.
return input.providerId === "codex"
? appendSyncedEffortVariants(
exclusiveMerged.map((model) => ({ ...model, owned_by: "codex" }))
)
: exclusiveMerged;
}
const syncedById = new Map(synced.map((model) => [model.id, model]));
const builtInModels = input.registryModels.map((model) => ({
...model,
...(input.providerId === "codex" ? syncedById.get(model.id) : {}),
source: "system",
}));
const registryIds = new Set(builtInModels.map((model) => model.id));
const syncedExtras = synced
.filter((model) => model.id && !registryIds.has(model.id))
.map((model) => ({
...model,
id: model.id,
name: model.name || model.id,
source: "imported",
}));
const normalizedCustom = custom.map((model) => ({
...model,
id: model.id,
name: model.name || model.id,
source: normalizeCustomSource(model.source),
}));
const merged = dedupeById(
mergeModelsWithCustomPrecedence([...builtInModels, ...syncedExtras], normalizedCustom)
);
return input.providerId === "codex"
? appendSyncedEffortVariants(merged.map((model) => ({ ...model, owned_by: "codex" })))
: merged;
}