@arnilo/prism-provider-zai 0.0.1 → 0.0.2

package/CHANGELOG.md

@@ -7,6 +7,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.0.2] - 2026-07-05
+
 ### Added
 
 - Added `LICENSE` (MIT) and `CHANGELOG.md`.

package/README.md

@@ -18,3 +18,8 @@ Security defaults:
 - No network calls during import, setup, build, or default tests.
 - No automatic environment, file, keychain, or shell credential lookup.
 - API keys are resolved per request from caller-supplied values or resolvers.
+
+Cache behavior:
+- `kind: "implicit"` — GLM context caching is automatic; no explicit cache payload sent regardless of cache options.
+- `prompt_tokens_details.cached_tokens`/`cache_write_tokens` map to `Usage.cacheReadTokens`/`cacheWriteTokens`.
+- Provider-owned headers (`content-type`, `authorization`) win over caller headers.

package/dist/models.js

@@ -6,12 +6,14 @@ export const zaiModels = [
         model: "glm-4.7",
         displayName: "GLM-4.7",
         limits: { contextWindow: 128_000, maxOutputTokens: 32_000 },
+        cache: { kind: "implicit" },
         compat: { thinking: true, reasoning_effort: "medium", tool_stream: true },
     }),
     defineZaiModel({
         model: "glm-4.5",
         displayName: "GLM-4.5",
         limits: { contextWindow: 128_000, maxOutputTokens: 32_000 },
+        cache: { kind: "implicit" },
         compat: { thinking: true, reasoning_effort: "medium", tool_stream: true },
     }),
 ];

package/dist/provider.js

@@ -14,7 +14,7 @@ export function createZaiProvider(options = {}) {
             try {
                 const response = await (options.fetch ?? fetch)(`${baseUrl}/chat/completions`, {
                     method: "POST",
-                    headers: { "content-type": "application/json", ...request.options?.headers, ...(token ? { authorization: `Bearer ${token}` } : {}) },
+                    headers: { ...request.options?.headers, "content-type": "application/json", ...(token ? { authorization: `Bearer ${token}` } : {}) },
                     body: JSON.stringify(zaiBody(request)),
                     signal: request.signal,
                 });
@@ -31,6 +31,7 @@ export function createZaiProvider(options = {}) {
     };
 }
 export function zaiBody(request) {
+    const { maxTokens, ...parameters } = request.model.parameters ?? {};
     return clean({
         model: request.model.model,
         messages: request.messages.map((message) => toMessage(message, request.model.capabilities ?? {})),
@@ -39,8 +40,8 @@ export function zaiBody(request) {
         tool_stream: zaiToolStream(request),
         thinking: zaiThinking(request),
         reasoning_effort: zaiReasoningEffort(request),
-        max_tokens: request.model.limits?.maxOutputTokens,
-        ...request.model.parameters,
+        ...parameters,
+        max_tokens: maxTokens ?? request.model.limits?.maxOutputTokens,
         ...request.options?.compat,
         ...request.options?.extra,
     });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@arnilo/prism-provider-zai",
-  "version": "0.0.1",
+  "version": "0.0.2",
   "description": "Z.AI provider package for Prism.",
   "type": "module",
   "main": "./dist/index.js",
@@ -25,7 +25,7 @@
     "pack:dry-run": "npm pack --dry-run"
   },
   "peerDependencies": {
-    "@arnilo/prism": "0.0.1"
+    "@arnilo/prism": "0.0.2"
   },
   "devDependencies": {
     "@arnilo/prism": "file:../.."