The model-control wave (user-decided design, 2026-09-12): ONE controller
for the conversation agents (prober/teacher are functionally one agent);
background sub-agents stay config-pinned. pi-ai owns the registry — the
shipped 366-model OpenRouter catalog (cost, context, modality,
thinkingLevelMap), getSupportedThinkingLevels/clampThinkingLevel for the
per-model thinking truth, and streamSimple's automatic clamp. No second
registry to drift out of sync.
Seams:
- config.ts: ReasoningEffort retired for pi's ThinkingLevel
(off|minimal|low|medium|high|xhigh|max). GLM roles go honest at "low" —
pi's catalog says GLM-5.3-flash accepts exactly low|high|max, and the old
"minimal" was silently clamped to low by streamSimple since the swap;
nothing changes behaviorally. visualMaker → qwen/qwen3.8-flash (fixes
the dead qwen/qwen3.8-max slug — the catalog only ships -0902; also the
user's cost pick, 0.15/0.47 per-1M).
- src/agent/models.ts: piEngineFor(role, settings?) — provider+id resolved
from the catalog, thinking level from settings ?? config, CLAMPED to the
model's supported list (a settings change between models can never send
an invalid level). PiEngine gains `level` (observability/tests).
- ChatRecord.settings {model?: {provider, id}, thinkingLevel?} — written
by PUT /api/chats/[id]/settings (replace-whole, no updatedAt churn),
read by chat-handler at turn composition: agents are stateless per turn,
so the change is next-turn-effective by construction. The turns log
still records the ACTUAL billed model (the honest witness).
- GET /api/models: the configured-provider catalog as client-safe DTOs
(core/contracts/models.ts) + config defaults; a second provider joins
the picker by registering it in piModels() — nothing else changes.
- PUT settings validation gate (src/server/model-catalog.ts): unknown
provider/model or unsupported level → 400 carrying the supported list,
never a silent clamp.
- scripts/sync-pricing.ts: regenerated PRICES/CONTEXT_LIMITS from pi-ai's
shipped openrouter.json (offline; same source as the engine + picker).
All 366 models priced; glm's stale 0.15/0.5 → pi's 0.075/0.25, limit
1310720 → 1048576 — cost is inferred at read time, so history re-prices.
- UI (components/model-picker.tsx): stats-bar right side becomes model +
thinking buttons; menus open ABOVE; fuzzy search over "provider id
name" (core/fuzzy.ts — pi's /model pattern); model rows show name/id/
price/context; the thinking menu lists exactly the chosen model's
supported levels (GLM: low|high|max — no dead options, no "off").
Model switch drops an incompatible pinned level (falls back to config,
which clamps honestly) rather than carrying a 400.
Tests (350/350, +15): catalog projection + defaults, the settings gate,
pi thinking truth pinned against upgrades (glm low|high|max; minimal→low;
off→low — GLM cannot disable thinking), the engine factory's
override/fallback/clamp/throw matrix, fuzzy ranking, store round-trip,
and the re-priced folds (usage, stats bar, activity dock).
Live smoke (disposable driver, deleted): catalog route 366 + defaults;
glm+minimal → 400 + supported; turn 1 billed qwen/qwen3.8-flash under
the override; mid-session switch flipped turn 2 to glm; clear-overrides
restored the config pin. ALL CHECKS PASS on real OpenRouter.
108 lines
4.2 KiB
TypeScript
108 lines
4.2 KiB
TypeScript
/**
|
|
* The model catalog's server surface (model-control wave, 2026-09-12):
|
|
* pi-ai's shipped OpenRouter catalog projected to client-safe DTOs for the
|
|
* picker (GET /api/models), plus the validation gate for per-chat settings
|
|
* (PUT /api/chats/[id]/settings).
|
|
*
|
|
* pi-ai owns ALL the data: the catalog (id/name/cost/context/modality/
|
|
* thinkingLevelMap) and the per-model thinking truth
|
|
* (getSupportedThinkingLevels / clampThinkingLevel). This module only
|
|
* projects and validates — no second registry to drift out of sync.
|
|
*
|
|
* Multi-provider future: `getModels()` is called per configured provider,
|
|
* so a second provider (pi-ai ships ~30) appears in the picker by adding
|
|
* it to piModels() in src/agent/models.ts — nothing here changes.
|
|
*/
|
|
|
|
import { getSupportedThinkingLevels, type Api, type Model } from "@earendil-works/pi-ai";
|
|
import type { ThinkingLevel } from "@earendil-works/pi-agent-core";
|
|
import { MODELS } from "../../config.ts";
|
|
import { piModels } from "../agent/models.ts";
|
|
import type { ChatSettings } from "../core/contracts/chat.ts";
|
|
import type {
|
|
CatalogDefaults,
|
|
ModelCatalogEntry,
|
|
ModelCatalogResponse,
|
|
} from "../core/contracts/models.ts";
|
|
|
|
/** The picker's client-safe model entry (contract: core/contracts/models.ts). */
|
|
export type ModelDto = ModelCatalogEntry;
|
|
|
|
/** The conversation agents' config pin — what a settings-less chat runs. */
|
|
export function catalogDefaults(): CatalogDefaults {
|
|
return {
|
|
model: { provider: "openrouter", id: MODELS.prober.model },
|
|
thinkingLevel: MODELS.prober.effort as ThinkingLevel,
|
|
};
|
|
}
|
|
|
|
/** Project one catalog model to the client DTO. */
|
|
export function toModelDto(model: Model<Api>): ModelDto {
|
|
return {
|
|
id: model.id,
|
|
name: model.name,
|
|
provider: model.provider,
|
|
reasoning: model.reasoning,
|
|
input: [...model.input],
|
|
supportedLevels: getSupportedThinkingLevels(model),
|
|
cost: {
|
|
input: model.cost.input,
|
|
output: model.cost.output,
|
|
cacheRead: model.cost.cacheRead,
|
|
cacheWrite: model.cost.cacheWrite,
|
|
},
|
|
contextWindow: model.contextWindow,
|
|
maxTokens: model.maxTokens,
|
|
};
|
|
}
|
|
|
|
/** The full picker payload: every model of every configured provider
|
|
* (sorted by provider then id — stable, diff-friendly) + the config
|
|
* defaults a settings-less chat runs. */
|
|
export function modelCatalog(): ModelCatalogResponse {
|
|
const models = piModels()
|
|
.getProviders()
|
|
.flatMap((p) => piModels().getModels(p.id))
|
|
.map(toModelDto)
|
|
.sort((a, b) => (a.provider === b.provider ? a.id.localeCompare(b.id) : a.provider.localeCompare(b.provider)));
|
|
return { models, defaults: catalogDefaults() };
|
|
}
|
|
|
|
/** One validated outcome for a settings write. */
|
|
export type SettingsValidation =
|
|
| { ok: true }
|
|
| { ok: false; error: string; supported?: ThinkingLevel[] };
|
|
|
|
/** Validate a picker write against the catalog (the settings route's gate).
|
|
* `settings` is replace-whole: absent fields mean "config default".
|
|
* - model: provider must be configured, id must exist in its catalog.
|
|
* - thinkingLevel: must be in the resolved model's supported list —
|
|
* the model the picker writes against (settings.model, else config's
|
|
* prober/teacher pin; one controller, so the two specs share it). */
|
|
export function validateChatSettings(settings: ChatSettings): SettingsValidation {
|
|
const providers = piModels().getProviders().map((p) => p.id);
|
|
if (settings.model) {
|
|
const { provider, id } = settings.model;
|
|
if (!providers.includes(provider)) {
|
|
return { ok: false, error: `Unknown provider "${provider}" — configured: ${providers.join(", ")}` };
|
|
}
|
|
if (!piModels().getModel(provider, id)) {
|
|
return { ok: false, error: `No model "${id}" in the ${provider} catalog` };
|
|
}
|
|
}
|
|
if (settings.thinkingLevel !== undefined) {
|
|
const provider = settings.model?.provider ?? "openrouter";
|
|
const id = settings.model?.id ?? MODELS.prober.model;
|
|
const model = piModels().getModel(provider, id);
|
|
if (!model) return { ok: false, error: `No model "${id}" in the ${provider} catalog` };
|
|
const supported = getSupportedThinkingLevels(model);
|
|
if (!supported.includes(settings.thinkingLevel)) {
|
|
return {
|
|
ok: false,
|
|
error: `${id} does not support thinking level "${settings.thinkingLevel}"`,
|
|
supported,
|
|
};
|
|
}
|
|
}
|
|
return { ok: true };
|
|
} |