@oh-my-pi/pi-agent-core 15.10.7 → 15.10.9

package/CHANGELOG.md

@@ -2,6 +2,14 @@
 
 ## [Unreleased]
 
+## [15.10.8] - 2026-06-09
+### Added
+
+- Added optional `fetch` overrides to `SummaryOptions` and `compact`/`generateSummary` so remote compaction can use custom HTTP clients
+- Added optional `fetch` option to `ProxyStreamOptions` to control the HTTP request used by `streamProxy`
+- Added optional `fetch` overrides to `requestOpenAiRemoteCompaction` and `requestRemoteCompaction` for injectable HTTP transport
+- Added the upstream provider that served a request (`AssistantMessage.upstreamProvider`, e.g. OpenRouter's routed provider) as a `pi.gen_ai.response.upstream_provider` chat-span telemetry attribute, alongside the existing response id and time-to-first-chunk.
+
 ## [15.10.5] - 2026-06-08
 ### Removed
 

package/dist/types/compaction/compaction.d.ts

@@ -4,7 +4,7 @@
  * Pure functions for compaction logic. The session manager handles I/O,
  * and after compaction the session is reloaded.
  */
-import { type MessageAttribution, type Model, type Usage } from "@oh-my-pi/pi-ai";
+import { type FetchImpl, type MessageAttribution, type Model, type Usage } from "@oh-my-pi/pi-ai";
 import { type AgentTelemetry } from "../telemetry";
 import { ThinkingLevel } from "../thinking";
 import type { AgentMessage, AgentTool } from "../types";
@@ -125,6 +125,8 @@ export interface SummaryOptions {
      * `resolveCompactionEffort` for the conversion contract.
      */
     thinkingLevel?: ThinkingLevel;
+    /** Optional fetch implementation threaded into remote compaction calls. */
+    fetch?: FetchImpl;
 }
 export declare function generateSummary(currentMessages: AgentMessage[], model: Model, reserveTokens: number, apiKey: string, signal?: AbortSignal, customInstructions?: string, previousSummary?: string, options?: SummaryOptions): Promise<string>;
 export interface HandoffOptions {

package/dist/types/compaction/openai.d.ts

@@ -11,7 +11,7 @@
  *   summarization endpoints that accept `{ systemPrompt, prompt }` and reply
  *   with `{ summary, shortSummary? }`.
  */
-import type { Message, Model } from "@oh-my-pi/pi-ai/types";
+import type { FetchImpl, Message, Model } from "@oh-my-pi/pi-ai/types";
 export declare const OPENAI_REMOTE_COMPACTION_PRESERVE_KEY = "openaiRemoteCompaction";
 export type OpenAiRemoteCompactionItem = {
     type: "compaction" | "compaction_summary";
@@ -54,5 +54,9 @@ export declare function withOpenAiRemoteCompactionPreserveData(preserveData: Rec
  *   encrypted reasoning we want to preserve.
  */
 export declare function buildOpenAiNativeHistory(messages: Message[], model: Model, previousReplacementHistory?: Array<Record<string, unknown>>): Array<Record<string, unknown>>;
-export declare function requestOpenAiRemoteCompaction(model: Model, apiKey: string, compactInput: Array<Record<string, unknown>>, instructions: string, signal?: AbortSignal): Promise<OpenAiRemoteCompactionResponse>;
-export declare function requestRemoteCompaction(endpoint: string, request: RemoteCompactionRequest, signal?: AbortSignal): Promise<RemoteCompactionResponse>;
+export declare function requestOpenAiRemoteCompaction(model: Model, apiKey: string, compactInput: Array<Record<string, unknown>>, instructions: string, signal?: AbortSignal, opts?: {
+    fetch?: FetchImpl;
+}): Promise<OpenAiRemoteCompactionResponse>;
+export declare function requestRemoteCompaction(endpoint: string, request: RemoteCompactionRequest, signal?: AbortSignal, opts?: {
+    fetch?: FetchImpl;
+}): Promise<RemoteCompactionResponse>;

package/dist/types/proxy.d.ts

@@ -2,7 +2,7 @@
  * Proxy stream function for apps that route LLM calls through a server.
  * The server manages auth and proxies requests to LLM providers.
  */
-import { type AssistantMessage, type AssistantMessageEvent, type Context, EventStream, type Model, type SimpleStreamOptions, type StopReason } from "@oh-my-pi/pi-ai";
+import { type AssistantMessage, type AssistantMessageEvent, type Context, EventStream, type FetchImpl, type Model, type SimpleStreamOptions, type StopReason } from "@oh-my-pi/pi-ai";
 export declare class ProxyMessageEventStream extends EventStream<AssistantMessageEvent, AssistantMessage> {
     constructor();
 }
@@ -60,6 +60,8 @@ export interface ProxyStreamOptions extends SimpleStreamOptions {
     authToken: string;
     /** Proxy server URL (e.g., "https://genai.example.com") */
     proxyUrl: string;
+    /** Optional fetch implementation; defaults to global fetch. */
+    fetch?: FetchImpl;
 }
 /**
  * Stream function that proxies through a server instead of calling LLM providers directly.

package/dist/types/telemetry.d.ts

@@ -87,6 +87,7 @@ export declare const enum PiGenAIAttr {
     RequestMessages = "pi.gen_ai.request.messages",
     ResponseText = "pi.gen_ai.response.text",
     ResponseToolCalls = "pi.gen_ai.response.tool_calls",
+    ResponseUpstreamProvider = "pi.gen_ai.response.upstream_provider",
     UsageTotalTokens = "pi.gen_ai.usage.total_tokens",
     UsageServerSideTools = "pi.gen_ai.usage.server_tool_requests",
     CostEstimatedUsd = "pi.gen_ai.cost.estimated_usd",

package/package.json

@@ -1,7 +1,7 @@
 {
 	"type": "module",
 	"name": "@oh-my-pi/pi-agent-core",
-	"version": "15.10.7",
+	"version": "15.10.9",
 	"description": "General-purpose agent with transport abstraction, state management, and attachment support",
 	"homepage": "https://omp.sh",
 	"author": "Can Boluk",
@@ -35,9 +35,9 @@
 		"fmt": "biome format --write ."
 	},
 	"dependencies": {
-		"@oh-my-pi/pi-ai": "15.10.7",
-		"@oh-my-pi/pi-natives": "15.10.7",
-		"@oh-my-pi/pi-utils": "15.10.7",
+		"@oh-my-pi/pi-ai": "15.10.9",
+		"@oh-my-pi/pi-natives": "15.10.9",
+		"@oh-my-pi/pi-utils": "15.10.9",
 		"@opentelemetry/api": "^1.9.1"
 	},
 	"devDependencies": {

package/src/compaction/compaction.ts

@@ -9,6 +9,7 @@ import {
 	type AssistantMessage,
 	clampThinkingLevelForModel,
 	Effort,
+	type FetchImpl,
 	type Message,
 	type MessageAttribution,
 	type Model,
@@ -595,6 +596,8 @@ export interface SummaryOptions {
 	 * `resolveCompactionEffort` for the conversion contract.
 	 */
 	thinkingLevel?: ThinkingLevel;
+	/** Optional fetch implementation threaded into remote compaction calls. */
+	fetch?: FetchImpl;
 }
 
 export async function generateSummary(
@@ -647,6 +650,7 @@ export async function generateSummary(
 				prompt: promptText,
 			},
 			signal,
+			{ fetch: options.fetch },
 		);
 		return remote.summary;
 	}
@@ -992,6 +996,7 @@ export async function compact(
 		// silently falls back to Effort.High — the same defect e07b47ee4 fixed
 		// at the call sites, leaked back in here. See resolveCompactionEffort.
 		thinkingLevel: options?.thinkingLevel,
+		fetch: options?.fetch,
 	};
 
 	let preserveData = withOpenAiRemoteCompactionPreserveData(previousPreserveData, undefined);
@@ -1015,6 +1020,7 @@ export async function compact(
 					remoteHistory,
 					summaryOptions.remoteInstructions ?? SUMMARIZATION_SYSTEM_PROMPT,
 					signal,
+					{ fetch: summaryOptions.fetch },
 				);
 				preserveData = withOpenAiRemoteCompactionPreserveData(previousPreserveData, remote);
 			} catch (err) {

package/src/compaction/openai.ts

@@ -20,7 +20,7 @@ import {
 } from "@oh-my-pi/pi-ai/providers/openai-codex/constants";
 import { parseTextSignature } from "@oh-my-pi/pi-ai/providers/openai-responses-shared";
 import { transformMessages } from "@oh-my-pi/pi-ai/providers/transform-messages";
-import type { AssistantMessage, Message, Model } from "@oh-my-pi/pi-ai/types";
+import type { AssistantMessage, FetchImpl, Message, Model } from "@oh-my-pi/pi-ai/types";
 import {
 	getOpenAIResponsesHistoryItems,
 	getOpenAIResponsesHistoryPayload,
@@ -428,6 +428,7 @@ export async function requestOpenAiRemoteCompaction(
 	compactInput: Array<Record<string, unknown>>,
 	instructions: string,
 	signal?: AbortSignal,
+	opts?: { fetch?: FetchImpl },
 ): Promise<OpenAiRemoteCompactionResponse> {
 	const endpoint = resolveOpenAiCompactEndpoint(model);
 	const request: OpenAiRemoteCompactionRequest = {
@@ -451,7 +452,7 @@ export async function requestOpenAiRemoteCompaction(
 		headers[OPENAI_HEADERS.ORIGINATOR] = OPENAI_HEADER_VALUES.ORIGINATOR_CODEX;
 	}
 
-	const response = await fetch(endpoint, {
+	const response = await (opts?.fetch ?? fetch)(endpoint, {
 		method: "POST",
 		headers,
 		body: JSON.stringify(request),
@@ -501,8 +502,9 @@ export async function requestRemoteCompaction(
 	endpoint: string,
 	request: RemoteCompactionRequest,
 	signal?: AbortSignal,
+	opts?: { fetch?: FetchImpl },
 ): Promise<RemoteCompactionResponse> {
-	const response = await fetch(endpoint, {
+	const response = await (opts?.fetch ?? fetch)(endpoint, {
 		method: "POST",
 		headers: { "content-type": "application/json" },
 		body: JSON.stringify(request),

package/src/proxy.ts

@@ -7,6 +7,7 @@ import {
 	type AssistantMessageEvent,
 	type Context,
 	EventStream,
+	type FetchImpl,
 	type Model,
 	type SimpleStreamOptions,
 	type StopReason,
@@ -61,6 +62,8 @@ export interface ProxyStreamOptions extends SimpleStreamOptions {
 	authToken: string;
 	/** Proxy server URL (e.g., "https://genai.example.com") */
 	proxyUrl: string;
+	/** Optional fetch implementation; defaults to global fetch. */
+	fetch?: FetchImpl;
 }
 
 /**
@@ -117,7 +120,7 @@ export function streamProxy(model: Model, context: Context, options: ProxyStream
 		}
 
 		try {
-			response = await fetch(`${options.proxyUrl}/api/stream`, {
+			response = await (options.fetch ?? fetch)(`${options.proxyUrl}/api/stream`, {
 				method: "POST",
 				headers: {
 					Authorization: `Bearer ${options.authToken}`,

package/src/telemetry.ts

@@ -133,6 +133,7 @@ export const enum PiGenAIAttr {
 	RequestMessages = "pi.gen_ai.request.messages",
 	ResponseText = "pi.gen_ai.response.text",
 	ResponseToolCalls = "pi.gen_ai.response.tool_calls",
+	ResponseUpstreamProvider = "pi.gen_ai.response.upstream_provider",
 	UsageTotalTokens = "pi.gen_ai.usage.total_tokens",
 	UsageServerSideTools = "pi.gen_ai.usage.server_tool_requests",
 	CostEstimatedUsd = "pi.gen_ai.cost.estimated_usd",
@@ -1189,6 +1190,9 @@ export function failChatSpan(
 function applyChatResponseAttributes(span: Span, message: AssistantMessage): void {
 	span.setAttribute(GenAIAttr.ResponseModel, message.model);
 	if (message.responseId) span.setAttribute(GenAIAttr.ResponseId, message.responseId);
+	if (message.upstreamProvider) {
+		span.setAttribute(PiGenAIAttr.ResponseUpstreamProvider, message.upstreamProvider);
+	}
 	if (message.ttft != null) span.setAttribute(GenAIAttr.ResponseTimeToFirstChunk, message.ttft / 1000);
 	const finishReason = mapStopReason(message.stopReason);
 	if (finishReason) span.setAttribute(GenAIAttr.ResponseFinishReasons, [finishReason]);