diff --git a/CHANGELOG.md b/CHANGELOG.md index 81128f9..f861b13 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,10 +2,8 @@ ## Unreleased -- Reformat `/commandcode-quota` output into a dashboard layout: credits remaining/used with a percentage, monthly/purchased/free sources, plan, month-to-date cost/requests/tokens, and API key name, while keeping the 5-hour/weekly usage windows and full-detail link. -- Optionally show an aggregate token count and API key name when the usage endpoints report them. -- Fix `/commandcode-quota` on Oh My Pi: OMP surfaces an unresolved `$COMMANDCODE_API_KEY` placeholder through the model registry, which previously caused a 401 on the quota endpoints. The command now filters placeholders and falls back to the env/auth-file key resolver, matching the stream path. -- Fix `/commandcode-quota` usage windows so the 5-hour and weekly limits actually render: the API reports `windowLimits` as a top-level sibling of `credits` (not nested inside it), and its `resetAt` is in milliseconds, not seconds. Both are now parsed correctly, giving real "resets in …" countdowns instead of omitting the section or showing a huge day count. +- Add `/commandcode-quota` with live credits, plan, usage totals, and rolling-limit diagnostics from Command Code's alpha usage endpoints. +- Add `zai-org/GLM-5.3` with its verified reasoning efforts and display pricing. - Prefer Command Code's Provider API (`/provider/v1/chat/completions` and `/provider/v1/messages`) and automatically fall back to the existing `/alpha/generate` transport only when the Provider API returns `403 upgrade_required` for a Go-plan account. - Remember the detected transport for the running process, re-detect it when credentials change, prevent stale in-flight requests from overwriting the new credential's transport, and never fall back for unrelated authentication, permission, rate-limit, network, or server failures. - Use Pi's native OpenAI- and Anthropic-compatible providers for Provider API streaming, including adaptive thinking for current reasoning-capable Claude models, while preserving the existing hardened generate transport, dynamic model discovery, offline cache, refresh/status commands, pricing, and OAuth credentials. @@ -13,6 +11,7 @@ - Add optional zero-data-retention headers through `COMMANDCODE_ZDR=1`. - Refresh GPT-5.6 Terra and Luna display prices after their temporary 50% promotion ended, and display the current DeepSeek V4 off-peak rates for its time-dependent pricing. - Add isolated live E2E profiles for separate Go-plan and Provider-API credentials, including an explicit selected-transport assertion and packed-package validation. +- Fix extension load failure on newer pi hosts that reject registering a custom API under a built-in name (`openai-completions`); register under `commandcode-custom` instead and restore the real wire API before native compat dispatch. ## 0.5.1 - 2026-08-11 diff --git a/README.md b/README.md index b140206..2de94fb 100644 --- a/README.md +++ b/README.md @@ -144,7 +144,7 @@ The following environment variables are intended for tests, local mocks, and com ## Image input -The provider advertises image input only for models marked with the `image` input modality in the official Command Code CLI model catalog. The capability snapshot currently follows `command-code@1.15.1`; unknown models default to text-only until their upstream metadata is reviewed. +The provider advertises image input only for models marked with the `image` input modality in the official Command Code CLI model catalog. The capability snapshot currently follows `command-code@1.32.1`; unknown models default to text-only until their upstream metadata is reviewed. For vision-capable models, Pi's native provider adapters forward image blocks from user messages and tool results using the documented OpenAI or Anthropic message schema. Unknown and text-only models remain marked text-only in Pi. diff --git a/index.ts b/index.ts index be7891f..1a7bd57 100644 --- a/index.ts +++ b/index.ts @@ -20,6 +20,7 @@ import { createStreamCommandCode } from "./src/core.ts" import { calculateCommandCodeCost } from "./src/cost.ts" import { pickCommandCodeApiKey } from "./src/converters.ts" import { + apiForModelId, baseUrlForModel, DEFAULT_MODELS_URL, DEFAULT_PROVIDER_API_BASE, @@ -63,7 +64,7 @@ function createProviderConfig( name: "Command Code", baseUrl: apiBase, apiKey: getConfiguredApiKey() ?? "$COMMANDCODE_API_KEY", - api: "openai-completions", + api: "commandcode-custom", streamSimple: streamCommandCode, headers, oauth: { @@ -75,7 +76,7 @@ function createProviderConfig( models: models.map((model) => ({ id: model.id, name: model.name, - api: model.api, + api: "commandcode-custom", baseUrl: baseUrlForModel(apiBase, model.api), reasoning: model.reasoning, ...(thinkingMetadataForModel(model.id) ?? {}), @@ -120,7 +121,12 @@ export default async function (pi: ExtensionAPI) { }) const transport = createCommandCodeTransportRouter({ createStream: () => new AssistantMessageEventStream(), - streamProvider: streamNativeProvider, + streamProvider: (model, context, options) => + streamNativeProvider( + { ...model, api: apiForModelId(model.id), compat: model.compatConfig ?? model.compat }, + context, + options, + ), streamGenerate, }) diff --git a/src/core.ts b/src/core.ts index f1898dd..2eb92f8 100644 --- a/src/core.ts +++ b/src/core.ts @@ -43,7 +43,7 @@ export * from "./overflow.ts" export * from "./types.ts" export const DEFAULT_API_BASE = "https://api.commandcode.ai" -export const COMMAND_CODE_CLI_VERSION = "1.15.1" +export const COMMAND_CODE_CLI_VERSION = "1.32.1" const DEFAULT_GENERATE_MAX_TOKENS = 64_000 const DEFAULT_MAX_RETRIES = 0 diff --git a/src/models.ts b/src/models.ts index 3f8c7e9..e102c00 100644 --- a/src/models.ts +++ b/src/models.ts @@ -12,7 +12,7 @@ export type CommandCodeApi = "openai-completions" | "anthropic-messages" export type CommandCodeInputType = "text" | "image" /** - * Model input modalities from the command-code@1.15.1 bundled catalog. + * Model input modalities from the command-code@1.32.1 bundled catalog. * Models omitted here remain text-only so newly discovered IDs never claim * image support without upstream evidence. */ @@ -74,7 +74,7 @@ type CommandCodeReasoningEffort = Exclude * Per-model reasoning efforts supported by Command Code's generate endpoint. * * The Provider API does not expose reasoning metadata. This is an exact - * snapshot of `reasoningEfforts` from the command-code@1.15.1 model catalog + * snapshot of `reasoningEfforts` from the command-code@1.32.1 model catalog * (`packages/shared/src/model-catalog.ts`, also published in the generated * `dist/bundled/command-code-knowledge/reference/models.md`). Models omitted * here let Command Code choose their reasoning depth, matching the CLI. @@ -103,6 +103,7 @@ export const MODEL_EFFORTS: Readonly> = { }, "moonshotai/Kimi-K2.6": { input: 0.95, output: 4, cacheRead: 0.16, cacheWrite: 0 }, "moonshotai/Kimi-K2.5": { input: 0.6, output: 3, cacheRead: 0.1, cacheWrite: 0 }, + "zai-org/GLM-5.3": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, "zai-org/GLM-5.2": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, "zai-org/GLM-5.2-Fast": { input: 3, output: 10.25, cacheRead: 0.5, cacheWrite: 0 }, "zai-org/GLM-5.1": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, diff --git a/tests/fixtures/commandcode-model-ids.json b/tests/fixtures/commandcode-model-ids.json index f0974cd..1b9898a 100644 --- a/tests/fixtures/commandcode-model-ids.json +++ b/tests/fixtures/commandcode-model-ids.json @@ -1,5 +1,5 @@ { - "fetchedAt": "2026-08-04T10:12:57.953Z", + "fetchedAt": "2026-08-22T21:19:37.782Z", "source": "https://api.commandcode.ai/provider/v1/models", "modelIds": [ "claude-sonnet-5", @@ -23,6 +23,7 @@ "moonshotai/Kimi-K2.7-Code-Highspeed", "moonshotai/Kimi-K2.6", "moonshotai/Kimi-K2.5", + "zai-org/GLM-5.3", "zai-org/GLM-5.2", "zai-org/GLM-5.2-Fast", "zai-org/GLM-5.1", diff --git a/tests/fixtures/commandcode-pricing.json b/tests/fixtures/commandcode-pricing.json index 6a508fd..5b2070c 100644 --- a/tests/fixtures/commandcode-pricing.json +++ b/tests/fixtures/commandcode-pricing.json @@ -1,5 +1,5 @@ { - "verifiedAt": "2026-08-20", + "verifiedAt": "2026-08-22", "source": "https://commandcode.ai/docs/resources/pricing-limits", "tierPolicy": "Use request-wide input tiers; the highest threshold exceeded by input plus cache tokens applies to the full request.", "tiers": { @@ -18,6 +18,7 @@ "moonshotai/Kimi-K2.7-Code-Highspeed": [1.9, 8, 0.38, 0], "moonshotai/Kimi-K2.6": [0.95, 4, 0.16, 0], "moonshotai/Kimi-K2.5": [0.6, 3, 0.1, 0], + "zai-org/GLM-5.3": [1.4, 4.4, 0.26, 0], "zai-org/GLM-5.2": [1.4, 4.4, 0.26, 0], "zai-org/GLM-5.2-Fast": [3, 10.25, 0.5, 0], "zai-org/GLM-5.1": [1.4, 4.4, 0.26, 0], diff --git a/tests/test-models.ts b/tests/test-models.ts index 3c7e33c..f10dd7e 100644 --- a/tests/test-models.ts +++ b/tests/test-models.ts @@ -100,7 +100,7 @@ describe("commandCodeModelsFromApiResponse()", () => { ) }) - it("matches command-code@1.15.1 image input capabilities", () => { + it("matches command-code@1.32.1 image input capabilities", () => { assert.deepEqual(inputModalitiesForModel("gpt-5.6-luna"), ["text", "image"]) assert.deepEqual(inputModalitiesForModel("meta/muse-spark-1.2"), ["text", "image"]) assert.deepEqual(inputModalitiesForModel("deepseek/deepseek-v4-pro"), ["text"]) @@ -123,7 +123,7 @@ describe("commandCodeModelsFromApiResponse()", () => { assert.equal(models[1]?.reasoning, false) }) - it("matches the exact command-code@1.15.1 reasoning effort catalog", () => { + it("matches the exact command-code@1.32.1 reasoning effort catalog", () => { assert.deepEqual(MODEL_EFFORTS, { "Qwen/Qwen3.8-Max": ["low", "medium", "xhigh"], "claude-fable-5": ["low", "medium", "high", "xhigh", "max"], @@ -147,6 +147,7 @@ describe("commandCodeModelsFromApiResponse()", () => { "google/gemini-3.6-flash": ["low", "medium", "high"], "sakana/fugu-ultra": ["high", "xhigh"], "xai/grok-4.5": ["low", "medium", "high"], + "zai-org/GLM-5.3": ["low", "high", "max"], "zai-org/GLM-5.2": ["high", "max"], }) }) diff --git a/tests/test-pricing.ts b/tests/test-pricing.ts index 6e17400..202d8aa 100644 --- a/tests/test-pricing.ts +++ b/tests/test-pricing.ts @@ -50,7 +50,7 @@ function assertCost( describe("MODEL_COSTS pricing overlay", () => { it("covers the current Command Code model catalog snapshot", () => { assert.equal(fixture.source, "https://api.commandcode.ai/provider/v1/models") - assert.match(fixture.fetchedAt, /^2026-08-04T/) + assert.match(fixture.fetchedAt, /^2026-08-22T/) const catalogIds = [...fixture.modelIds].sort() const pricedIds = Object.keys(MODEL_COSTS).sort() @@ -169,7 +169,7 @@ describe("MODEL_COSTS pricing overlay", () => { it("tracks pricing provenance", () => { assert.equal(PRICING_SOURCE_URL, "https://commandcode.ai/docs/resources/pricing-limits") - assert.equal(PRICING_LAST_VERIFIED, "2026-08-20") + assert.equal(PRICING_LAST_VERIFIED, "2026-08-22") }) it("fails once temporary pricing needs review", () => { diff --git a/tests/test-stream.ts b/tests/test-stream.ts index e30ac49..a429798 100644 --- a/tests/test-stream.ts +++ b/tests/test-stream.ts @@ -488,7 +488,7 @@ describe("streamCommandCode — request serialization", () => { const headers = server.lastRequestHeaders() assert.equal(headers.authorization, "Bearer mock-key") - assert.equal(headers["x-command-code-version"], "1.15.1") + assert.equal(headers["x-command-code-version"], "1.32.1") assert.equal(headers["x-project-slug"], "repo") assert.equal(headers["x-taste-learning"], "true") assert.equal(headers["x-co-flag"], "false")