@steipete/summarize 0.7.1 → 0.8.1

package/dist/esm/run/flows/url/summary.js

@@ -1,6 +1,8 @@
 import { countTokens } from 'gpt-tokenizer';
 import { render as renderMarkdownAnsi } from 'markdansi';
+import { buildLanguageKey, buildLengthKey, buildPromptHash, buildSummaryCacheKey, hashString, normalizeContentForHash, } from '../../../cache.js';
 import { formatOutputLanguageForJson } from '../../../language.js';
+import { parseGatewayStyleModelId } from '../../../llm/model-id.js';
 import { buildAutoModelAttempts } from '../../../model-auto.js';
 import { buildLinkSummaryPrompt } from '../../../prompts/index.js';
 import { parseCliUserModelId } from '../../env.js';
@@ -10,7 +12,7 @@ import { prepareMarkdownForTerminal } from '../../markdown.js';
 import { runModelAttempts } from '../../model-attempts.js';
 import { buildOpenRouterNoAllowedProvidersMessage } from '../../openrouter.js';
 import { isRichTty, markdownRenderWidth, supportsColor } from '../../terminal.js';
-export function buildUrlPrompt({ extracted, outputLanguage, lengthArg, }) {
+export function buildUrlPrompt({ extracted, outputLanguage, lengthArg, promptOverride, lengthInstruction, languageInstruction, }) {
     const isYouTube = extracted.siteName === 'YouTube';
     return buildLinkSummaryPrompt({
         url: extracted.url,
@@ -18,12 +20,15 @@ export function buildUrlPrompt({ extracted, outputLanguage, lengthArg, }) {
         siteName: extracted.siteName,
         description: extracted.description,
         content: extracted.content,
-        truncated: false,
+        truncated: extracted.truncated,
         hasTranscript: isYouTube ||
             (extracted.transcriptSource !== null && extracted.transcriptSource !== 'unavailable'),
         summaryLength: lengthArg.kind === 'preset' ? lengthArg.preset : { maxCharacters: lengthArg.maxCharacters },
         outputLanguage,
         shares: [],
+        promptOverride: promptOverride ?? null,
+        lengthInstruction: lengthInstruction ?? null,
+        languageInstruction: languageInstruction ?? null,
     });
 }
 const buildFinishExtras = ({ extracted, metricsDetailed, transcriptionCostLabel, }) => {
@@ -47,315 +52,386 @@ const pickModelForFinishLine = (llmCalls, fallback) => {
         (llmCalls.length > 0 ? (llmCalls[llmCalls.length - 1]?.model ?? null) : null) ??
         fallback);
 };
+const buildModelMetaFromAttempt = (attempt) => {
+    if (attempt.transport === 'cli') {
+        return { provider: 'cli', canonical: attempt.userModelId };
+    }
+    const parsed = parseGatewayStyleModelId(attempt.llmModelId ?? attempt.userModelId);
+    const canonical = attempt.userModelId.toLowerCase().startsWith('openrouter/')
+        ? attempt.userModelId
+        : parsed.canonical;
+    return { provider: parsed.provider, canonical };
+};
 export async function outputExtractedUrl({ ctx, url, extracted, extractionUi, prompt, effectiveMarkdownMode, transcriptionCostLabel, }) {
-    ctx.clearProgressForStdout();
+    const { io, flags, model, hooks } = ctx;
+    hooks.clearProgressForStdout();
     const finishLabel = buildExtractFinishLabel({
         extracted: { diagnostics: extracted.diagnostics },
-        format: ctx.format,
+        format: flags.format,
         markdownMode: effectiveMarkdownMode,
-        hasMarkdownLlmCall: ctx.llmCalls.some((call) => call.purpose === 'markdown'),
+        hasMarkdownLlmCall: model.llmCalls.some((call) => call.purpose === 'markdown'),
     });
-    const finishModel = pickModelForFinishLine(ctx.llmCalls, null);
-    if (ctx.json) {
-        const finishReport = ctx.shouldComputeReport ? await ctx.buildReport() : null;
+    const finishModel = pickModelForFinishLine(model.llmCalls, null);
+    if (flags.json) {
+        const finishReport = flags.shouldComputeReport ? await hooks.buildReport() : null;
         const payload = {
             input: {
                 kind: 'url',
                 url,
-                timeoutMs: ctx.timeoutMs,
-                youtube: ctx.youtubeMode,
-                firecrawl: ctx.firecrawlMode,
-                format: ctx.format,
+                timeoutMs: flags.timeoutMs,
+                youtube: flags.youtubeMode,
+                firecrawl: flags.firecrawlMode,
+                format: flags.format,
                 markdown: effectiveMarkdownMode,
-                length: ctx.lengthArg.kind === 'preset'
-                    ? { kind: 'preset', preset: ctx.lengthArg.preset }
-                    : { kind: 'chars', maxCharacters: ctx.lengthArg.maxCharacters },
-                maxOutputTokens: ctx.maxOutputTokensArg,
-                model: ctx.requestedModelLabel,
-                language: formatOutputLanguageForJson(ctx.outputLanguage),
+                length: flags.lengthArg.kind === 'preset'
+                    ? { kind: 'preset', preset: flags.lengthArg.preset }
+                    : { kind: 'chars', maxCharacters: flags.lengthArg.maxCharacters },
+                maxOutputTokens: flags.maxOutputTokensArg,
+                model: model.requestedModelLabel,
+                language: formatOutputLanguageForJson(flags.outputLanguage),
             },
             env: {
-                hasXaiKey: Boolean(ctx.apiStatus.xaiApiKey),
-                hasOpenAIKey: Boolean(ctx.apiStatus.apiKey),
-                hasOpenRouterKey: Boolean(ctx.apiStatus.openrouterApiKey),
-                hasApifyToken: Boolean(ctx.apiStatus.apifyToken),
-                hasFirecrawlKey: ctx.apiStatus.firecrawlConfigured,
-                hasGoogleKey: ctx.apiStatus.googleConfigured,
-                hasAnthropicKey: ctx.apiStatus.anthropicConfigured,
+                hasXaiKey: Boolean(model.apiStatus.xaiApiKey),
+                hasOpenAIKey: Boolean(model.apiStatus.apiKey),
+                hasOpenRouterKey: Boolean(model.apiStatus.openrouterApiKey),
+                hasApifyToken: Boolean(model.apiStatus.apifyToken),
+                hasFirecrawlKey: model.apiStatus.firecrawlConfigured,
+                hasGoogleKey: model.apiStatus.googleConfigured,
+                hasAnthropicKey: model.apiStatus.anthropicConfigured,
             },
             extracted,
             prompt,
             llm: null,
-            metrics: ctx.metricsEnabled ? finishReport : null,
+            metrics: flags.metricsEnabled ? finishReport : null,
             summary: null,
         };
-        ctx.stdout.write(`${JSON.stringify(payload, null, 2)}\n`);
-        if (ctx.metricsEnabled && finishReport) {
-            const costUsd = await ctx.estimateCostUsd();
+        io.stdout.write(`${JSON.stringify(payload, null, 2)}\n`);
+        if (flags.metricsEnabled && finishReport) {
+            const costUsd = await hooks.estimateCostUsd();
             writeFinishLine({
-                stderr: ctx.stderr,
-                elapsedMs: Date.now() - ctx.runStartedAtMs,
+                stderr: io.stderr,
+                elapsedMs: Date.now() - flags.runStartedAtMs,
                 label: finishLabel,
                 model: finishModel,
                 report: finishReport,
                 costUsd,
-                detailed: ctx.metricsDetailed,
+                detailed: flags.metricsDetailed,
                 extraParts: buildFinishExtras({
                     extracted,
-                    metricsDetailed: ctx.metricsDetailed,
+                    metricsDetailed: flags.metricsDetailed,
                     transcriptionCostLabel,
                 }),
-                color: ctx.verboseColor,
+                color: flags.verboseColor,
             });
         }
         return;
     }
-    const renderedExtract = ctx.format === 'markdown' && !ctx.plain && isRichTty(ctx.stdout)
+    const renderedExtract = flags.format === 'markdown' && !flags.plain && isRichTty(io.stdout)
         ? renderMarkdownAnsi(prepareMarkdownForTerminal(extracted.content), {
-            width: markdownRenderWidth(ctx.stdout, ctx.env),
+            width: markdownRenderWidth(io.stdout, io.env),
             wrap: true,
-            color: supportsColor(ctx.stdout, ctx.envForRun),
+            color: supportsColor(io.stdout, io.envForRun),
             hyperlinks: true,
         })
         : extracted.content;
-    if (ctx.format === 'markdown' && !ctx.plain && isRichTty(ctx.stdout)) {
-        ctx.stdout.write(`\n${renderedExtract.replace(/^\n+/, '')}`);
+    if (flags.format === 'markdown' && !flags.plain && isRichTty(io.stdout)) {
+        io.stdout.write(`\n${renderedExtract.replace(/^\n+/, '')}`);
     }
     else {
-        ctx.stdout.write(renderedExtract);
+        io.stdout.write(renderedExtract);
     }
     if (!renderedExtract.endsWith('\n')) {
-        ctx.stdout.write('\n');
+        io.stdout.write('\n');
     }
-    ctx.writeViaFooter(extractionUi.footerParts);
-    const report = ctx.shouldComputeReport ? await ctx.buildReport() : null;
-    if (ctx.metricsEnabled && report) {
-        const costUsd = await ctx.estimateCostUsd();
+    hooks.writeViaFooter(extractionUi.footerParts);
+    const report = flags.shouldComputeReport ? await hooks.buildReport() : null;
+    if (flags.metricsEnabled && report) {
+        const costUsd = await hooks.estimateCostUsd();
         writeFinishLine({
-            stderr: ctx.stderr,
-            elapsedMs: Date.now() - ctx.runStartedAtMs,
+            stderr: io.stderr,
+            elapsedMs: Date.now() - flags.runStartedAtMs,
             label: finishLabel,
             model: finishModel,
             report,
             costUsd,
-            detailed: ctx.metricsDetailed,
+            detailed: flags.metricsDetailed,
             extraParts: buildFinishExtras({
                 extracted,
-                metricsDetailed: ctx.metricsDetailed,
+                metricsDetailed: flags.metricsDetailed,
                 transcriptionCostLabel,
             }),
-            color: ctx.verboseColor,
+            color: flags.verboseColor,
         });
     }
 }
 export async function summarizeExtractedUrl({ ctx, url, extracted, extractionUi, prompt, effectiveMarkdownMode, transcriptionCostLabel, onModelChosen, }) {
+    const { io, flags, model, cache: cacheState, hooks } = ctx;
     const promptTokens = countTokens(prompt);
     const kindForAuto = extracted.siteName === 'YouTube' ? 'youtube' : 'website';
     const attempts = await (async () => {
-        if (ctx.isFallbackModel) {
-            const catalog = await ctx.getLiteLlmCatalog();
+        if (model.isFallbackModel) {
+            const catalog = await model.getLiteLlmCatalog();
             const list = buildAutoModelAttempts({
                 kind: kindForAuto,
                 promptTokens,
-                desiredOutputTokens: ctx.desiredOutputTokens,
+                desiredOutputTokens: model.desiredOutputTokens,
                 requiresVideoUnderstanding: false,
-                env: ctx.envForAuto,
-                config: ctx.configForModelSelection,
+                env: model.envForAuto,
+                config: model.configForModelSelection,
                 catalog,
                 openrouterProvidersFromEnv: null,
-                cliAvailability: ctx.cliAvailability,
+                cliAvailability: model.cliAvailability,
             });
-            if (ctx.verbose) {
+            if (flags.verbose) {
                 for (const attempt of list.slice(0, 8)) {
-                    writeVerbose(ctx.stderr, ctx.verbose, `auto candidate ${attempt.debug}`, ctx.verboseColor);
+                    writeVerbose(io.stderr, flags.verbose, `auto candidate ${attempt.debug}`, flags.verboseColor);
                 }
             }
             return list.map((attempt) => {
                 if (attempt.transport !== 'cli')
-                    return ctx.summaryEngine.applyZaiOverrides(attempt);
+                    return model.summaryEngine.applyZaiOverrides(attempt);
                 const parsed = parseCliUserModelId(attempt.userModelId);
                 return { ...attempt, cliProvider: parsed.provider, cliModel: parsed.model };
             });
         }
         /* v8 ignore next */
-        if (!ctx.fixedModelSpec) {
+        if (!model.fixedModelSpec) {
             throw new Error('Internal error: missing fixed model spec');
         }
-        if (ctx.fixedModelSpec.transport === 'cli') {
+        if (model.fixedModelSpec.transport === 'cli') {
             return [
                 {
                     transport: 'cli',
-                    userModelId: ctx.fixedModelSpec.userModelId,
+                    userModelId: model.fixedModelSpec.userModelId,
                     llmModelId: null,
-                    cliProvider: ctx.fixedModelSpec.cliProvider,
-                    cliModel: ctx.fixedModelSpec.cliModel,
+                    cliProvider: model.fixedModelSpec.cliProvider,
+                    cliModel: model.fixedModelSpec.cliModel,
                     openrouterProviders: null,
                     forceOpenRouter: false,
-                    requiredEnv: ctx.fixedModelSpec.requiredEnv,
+                    requiredEnv: model.fixedModelSpec.requiredEnv,
                 },
             ];
         }
-        const openaiOverrides = ctx.fixedModelSpec.requiredEnv === 'Z_AI_API_KEY'
+        const openaiOverrides = model.fixedModelSpec.requiredEnv === 'Z_AI_API_KEY'
             ? {
-                openaiApiKeyOverride: ctx.apiStatus.zaiApiKey,
-                openaiBaseUrlOverride: ctx.apiStatus.zaiBaseUrl,
+                openaiApiKeyOverride: model.apiStatus.zaiApiKey,
+                openaiBaseUrlOverride: model.apiStatus.zaiBaseUrl,
                 forceChatCompletions: true,
             }
             : {};
         return [
             {
-                transport: ctx.fixedModelSpec.transport === 'openrouter' ? 'openrouter' : 'native',
-                userModelId: ctx.fixedModelSpec.userModelId,
-                llmModelId: ctx.fixedModelSpec.llmModelId,
-                openrouterProviders: ctx.fixedModelSpec.openrouterProviders,
-                forceOpenRouter: ctx.fixedModelSpec.forceOpenRouter,
-                requiredEnv: ctx.fixedModelSpec.requiredEnv,
+                transport: model.fixedModelSpec.transport === 'openrouter' ? 'openrouter' : 'native',
+                userModelId: model.fixedModelSpec.userModelId,
+                llmModelId: model.fixedModelSpec.llmModelId,
+                openrouterProviders: model.fixedModelSpec.openrouterProviders,
+                forceOpenRouter: model.fixedModelSpec.forceOpenRouter,
+                requiredEnv: model.fixedModelSpec.requiredEnv,
                 ...openaiOverrides,
             },
         ];
     })();
-    const attemptOutcome = await runModelAttempts({
-        attempts,
-        isFallbackModel: ctx.isFallbackModel,
-        isNamedModelSelection: ctx.isNamedModelSelection,
-        envHasKeyFor: ctx.summaryEngine.envHasKeyFor,
-        formatMissingModelError: ctx.summaryEngine.formatMissingModelError,
-        onAutoSkip: (attempt) => {
-            writeVerbose(ctx.stderr, ctx.verbose, `auto skip ${attempt.userModelId}: missing ${attempt.requiredEnv}`, ctx.verboseColor);
-        },
-        onAutoFailure: (attempt, error) => {
-            writeVerbose(ctx.stderr, ctx.verbose, `auto failed ${attempt.userModelId}: ${error instanceof Error ? error.message : String(error)}`, ctx.verboseColor);
-        },
-        onFixedModelError: (_attempt, error) => {
-            throw error;
-        },
-        runAttempt: (attempt) => ctx.summaryEngine.runSummaryAttempt({
-            attempt,
-            prompt,
-            allowStreaming: ctx.streamingEnabled,
-            onModelChosen: onModelChosen ?? null,
-        }),
-    });
-    const summaryResult = attemptOutcome.result;
-    const usedAttempt = attemptOutcome.usedAttempt;
-    const { lastError, missingRequiredEnvs, sawOpenRouterNoAllowedProviders } = attemptOutcome;
+    const cacheStore = cacheState.mode === 'default' ? cacheState.store : null;
+    const contentHash = cacheStore ? hashString(normalizeContentForHash(extracted.content)) : null;
+    const promptHash = cacheStore ? buildPromptHash(prompt) : null;
+    const lengthKey = buildLengthKey(flags.lengthArg);
+    const languageKey = buildLanguageKey(flags.outputLanguage);
+    let summaryResult = null;
+    let usedAttempt = null;
+    let summaryFromCache = false;
+    let cacheChecked = false;
+    if (cacheStore && contentHash && promptHash) {
+        cacheChecked = true;
+        for (const attempt of attempts) {
+            if (!model.summaryEngine.envHasKeyFor(attempt.requiredEnv))
+                continue;
+            const key = buildSummaryCacheKey({
+                contentHash,
+                promptHash,
+                model: attempt.userModelId,
+                lengthKey,
+                languageKey,
+            });
+            const cached = cacheStore.getText('summary', key);
+            if (!cached)
+                continue;
+            writeVerbose(io.stderr, flags.verbose, 'cache hit summary', flags.verboseColor);
+            onModelChosen?.(attempt.userModelId);
+            summaryResult = {
+                summary: cached,
+                summaryAlreadyPrinted: false,
+                modelMeta: buildModelMetaFromAttempt(attempt),
+                maxOutputTokensForCall: null,
+            };
+            usedAttempt = attempt;
+            summaryFromCache = true;
+            break;
+        }
+    }
+    if (cacheChecked && !summaryFromCache) {
+        writeVerbose(io.stderr, flags.verbose, 'cache miss summary', flags.verboseColor);
+    }
+    ctx.hooks.onSummaryCached?.(summaryFromCache);
+    let lastError = null;
+    let missingRequiredEnvs = new Set();
+    let sawOpenRouterNoAllowedProviders = false;
+    if (!summaryResult || !usedAttempt) {
+        const attemptOutcome = await runModelAttempts({
+            attempts,
+            isFallbackModel: model.isFallbackModel,
+            isNamedModelSelection: model.isNamedModelSelection,
+            envHasKeyFor: model.summaryEngine.envHasKeyFor,
+            formatMissingModelError: model.summaryEngine.formatMissingModelError,
+            onAutoSkip: (attempt) => {
+                writeVerbose(io.stderr, flags.verbose, `auto skip ${attempt.userModelId}: missing ${attempt.requiredEnv}`, flags.verboseColor);
+            },
+            onAutoFailure: (attempt, error) => {
+                writeVerbose(io.stderr, flags.verbose, `auto failed ${attempt.userModelId}: ${error instanceof Error ? error.message : String(error)}`, flags.verboseColor);
+            },
+            onFixedModelError: (_attempt, error) => {
+                throw error;
+            },
+            runAttempt: (attempt) => model.summaryEngine.runSummaryAttempt({
+                attempt,
+                prompt,
+                allowStreaming: flags.streamingEnabled,
+                onModelChosen: onModelChosen ?? null,
+            }),
+        });
+        summaryResult = attemptOutcome.result;
+        usedAttempt = attemptOutcome.usedAttempt;
+        lastError = attemptOutcome.lastError;
+        missingRequiredEnvs = attemptOutcome.missingRequiredEnvs;
+        sawOpenRouterNoAllowedProviders = attemptOutcome.sawOpenRouterNoAllowedProviders;
+    }
     if (!summaryResult || !usedAttempt) {
         // Auto mode: surface raw extracted content when no model can run.
         const withFreeTip = (message) => {
-            if (!ctx.isNamedModelSelection || !ctx.wantsFreeNamedModel)
+            if (!model.isNamedModelSelection || !model.wantsFreeNamedModel)
                 return message;
             return (`${message}\n` +
                 `Tip: run "summarize refresh-free" to refresh the free model candidates (writes ~/.summarize/config.json).`);
         };
-        if (ctx.isNamedModelSelection) {
+        if (model.isNamedModelSelection) {
             if (lastError === null && missingRequiredEnvs.size > 0) {
-                throw new Error(withFreeTip(`Missing ${Array.from(missingRequiredEnvs).sort().join(', ')} for --model ${ctx.requestedModelInput}.`));
+                throw new Error(withFreeTip(`Missing ${Array.from(missingRequiredEnvs).sort().join(', ')} for --model ${model.requestedModelInput}.`));
             }
             if (lastError instanceof Error) {
                 if (sawOpenRouterNoAllowedProviders) {
                     const message = await buildOpenRouterNoAllowedProvidersMessage({
                         attempts,
-                        fetchImpl: ctx.trackedFetch,
-                        timeoutMs: ctx.timeoutMs,
+                        fetchImpl: io.fetch,
+                        timeoutMs: flags.timeoutMs,
                     });
                     throw new Error(withFreeTip(message), { cause: lastError });
                 }
                 throw new Error(withFreeTip(lastError.message), { cause: lastError });
             }
-            throw new Error(withFreeTip(`No model available for --model ${ctx.requestedModelInput}`));
+            throw new Error(withFreeTip(`No model available for --model ${model.requestedModelInput}`));
         }
-        ctx.clearProgressForStdout();
-        if (ctx.json) {
-            const finishReport = ctx.shouldComputeReport ? await ctx.buildReport() : null;
-            const finishModel = pickModelForFinishLine(ctx.llmCalls, null);
+        hooks.clearProgressForStdout();
+        if (flags.json) {
+            const finishReport = flags.shouldComputeReport ? await hooks.buildReport() : null;
+            const finishModel = pickModelForFinishLine(model.llmCalls, null);
             const payload = {
                 input: {
                     kind: 'url',
                     url,
-                    timeoutMs: ctx.timeoutMs,
-                    youtube: ctx.youtubeMode,
-                    firecrawl: ctx.firecrawlMode,
-                    format: ctx.format,
+                    timeoutMs: flags.timeoutMs,
+                    youtube: flags.youtubeMode,
+                    firecrawl: flags.firecrawlMode,
+                    format: flags.format,
                     markdown: effectiveMarkdownMode,
-                    length: ctx.lengthArg.kind === 'preset'
-                        ? { kind: 'preset', preset: ctx.lengthArg.preset }
-                        : { kind: 'chars', maxCharacters: ctx.lengthArg.maxCharacters },
-                    maxOutputTokens: ctx.maxOutputTokensArg,
-                    model: ctx.requestedModelLabel,
-                    language: formatOutputLanguageForJson(ctx.outputLanguage),
+                    length: flags.lengthArg.kind === 'preset'
+                        ? { kind: 'preset', preset: flags.lengthArg.preset }
+                        : { kind: 'chars', maxCharacters: flags.lengthArg.maxCharacters },
+                    maxOutputTokens: flags.maxOutputTokensArg,
+                    model: model.requestedModelLabel,
+                    language: formatOutputLanguageForJson(flags.outputLanguage),
                 },
                 env: {
-                    hasXaiKey: Boolean(ctx.apiStatus.xaiApiKey),
-                    hasOpenAIKey: Boolean(ctx.apiStatus.apiKey),
-                    hasOpenRouterKey: Boolean(ctx.apiStatus.openrouterApiKey),
-                    hasApifyToken: Boolean(ctx.apiStatus.apifyToken),
-                    hasFirecrawlKey: ctx.apiStatus.firecrawlConfigured,
-                    hasGoogleKey: ctx.apiStatus.googleConfigured,
-                    hasAnthropicKey: ctx.apiStatus.anthropicConfigured,
+                    hasXaiKey: Boolean(model.apiStatus.xaiApiKey),
+                    hasOpenAIKey: Boolean(model.apiStatus.apiKey),
+                    hasOpenRouterKey: Boolean(model.apiStatus.openrouterApiKey),
+                    hasApifyToken: Boolean(model.apiStatus.apifyToken),
+                    hasFirecrawlKey: model.apiStatus.firecrawlConfigured,
+                    hasGoogleKey: model.apiStatus.googleConfigured,
+                    hasAnthropicKey: model.apiStatus.anthropicConfigured,
                 },
                 extracted,
                 prompt,
                 llm: null,
-                metrics: ctx.metricsEnabled ? finishReport : null,
+                metrics: flags.metricsEnabled ? finishReport : null,
                 summary: extracted.content,
             };
-            ctx.stdout.write(`${JSON.stringify(payload, null, 2)}\n`);
-            if (ctx.metricsEnabled && finishReport) {
-                const costUsd = await ctx.estimateCostUsd();
+            io.stdout.write(`${JSON.stringify(payload, null, 2)}\n`);
+            if (flags.metricsEnabled && finishReport) {
+                const costUsd = await hooks.estimateCostUsd();
                 writeFinishLine({
-                    stderr: ctx.stderr,
-                    elapsedMs: Date.now() - ctx.runStartedAtMs,
+                    stderr: io.stderr,
+                    elapsedMs: Date.now() - flags.runStartedAtMs,
                     label: extractionUi.finishSourceLabel,
                     model: finishModel,
                     report: finishReport,
                     costUsd,
-                    detailed: ctx.metricsDetailed,
+                    detailed: flags.metricsDetailed,
                     extraParts: buildFinishExtras({
                         extracted,
-                        metricsDetailed: ctx.metricsDetailed,
+                        metricsDetailed: flags.metricsDetailed,
                         transcriptionCostLabel,
                     }),
-                    color: ctx.verboseColor,
+                    color: flags.verboseColor,
                 });
             }
             return;
         }
-        ctx.stdout.write(`${extracted.content}\n`);
+        io.stdout.write(`${extracted.content}\n`);
         if (extractionUi.footerParts.length > 0) {
-            ctx.writeViaFooter([...extractionUi.footerParts, 'no model']);
+            hooks.writeViaFooter([...extractionUi.footerParts, 'no model']);
         }
-        if (lastError instanceof Error && ctx.verbose) {
-            writeVerbose(ctx.stderr, ctx.verbose, `auto failed all models: ${lastError.message}`, ctx.verboseColor);
+        if (lastError instanceof Error && flags.verbose) {
+            writeVerbose(io.stderr, flags.verbose, `auto failed all models: ${lastError.message}`, flags.verboseColor);
         }
         return;
     }
+    if (!summaryFromCache && cacheStore && contentHash && promptHash) {
+        const key = buildSummaryCacheKey({
+            contentHash,
+            promptHash,
+            model: usedAttempt.userModelId,
+            lengthKey,
+            languageKey,
+        });
+        cacheStore.setText('summary', key, summaryResult.summary, cacheState.ttlMs);
+        writeVerbose(io.stderr, flags.verbose, 'cache write summary', flags.verboseColor);
+    }
     const { summary, summaryAlreadyPrinted, modelMeta, maxOutputTokensForCall } = summaryResult;
-    if (ctx.json) {
-        const finishReport = ctx.shouldComputeReport ? await ctx.buildReport() : null;
+    if (flags.json) {
+        const finishReport = flags.shouldComputeReport ? await hooks.buildReport() : null;
         const payload = {
             input: {
                 kind: 'url',
                 url,
-                timeoutMs: ctx.timeoutMs,
-                youtube: ctx.youtubeMode,
-                firecrawl: ctx.firecrawlMode,
-                format: ctx.format,
+                timeoutMs: flags.timeoutMs,
+                youtube: flags.youtubeMode,
+                firecrawl: flags.firecrawlMode,
+                format: flags.format,
                 markdown: effectiveMarkdownMode,
-                length: ctx.lengthArg.kind === 'preset'
-                    ? { kind: 'preset', preset: ctx.lengthArg.preset }
-                    : { kind: 'chars', maxCharacters: ctx.lengthArg.maxCharacters },
-                maxOutputTokens: ctx.maxOutputTokensArg,
-                model: ctx.requestedModelLabel,
-                language: formatOutputLanguageForJson(ctx.outputLanguage),
+                length: flags.lengthArg.kind === 'preset'
+                    ? { kind: 'preset', preset: flags.lengthArg.preset }
+                    : { kind: 'chars', maxCharacters: flags.lengthArg.maxCharacters },
+                maxOutputTokens: flags.maxOutputTokensArg,
+                model: model.requestedModelLabel,
+                language: formatOutputLanguageForJson(flags.outputLanguage),
             },
             env: {
-                hasXaiKey: Boolean(ctx.apiStatus.xaiApiKey),
-                hasOpenAIKey: Boolean(ctx.apiStatus.apiKey),
-                hasOpenRouterKey: Boolean(ctx.apiStatus.openrouterApiKey),
-                hasApifyToken: Boolean(ctx.apiStatus.apifyToken),
-                hasFirecrawlKey: ctx.apiStatus.firecrawlConfigured,
-                hasGoogleKey: ctx.apiStatus.googleConfigured,
-                hasAnthropicKey: ctx.apiStatus.anthropicConfigured,
+                hasXaiKey: Boolean(model.apiStatus.xaiApiKey),
+                hasOpenAIKey: Boolean(model.apiStatus.apiKey),
+                hasOpenRouterKey: Boolean(model.apiStatus.openrouterApiKey),
+                hasApifyToken: Boolean(model.apiStatus.apifyToken),
+                hasFirecrawlKey: model.apiStatus.firecrawlConfigured,
+                hasGoogleKey: model.apiStatus.googleConfigured,
+                hasAnthropicKey: model.apiStatus.anthropicConfigured,
             },
             extracted,
             prompt,
@@ -365,69 +441,71 @@ export async function summarizeExtractedUrl({ ctx, url, extracted, extractionUi,
                 maxCompletionTokens: maxOutputTokensForCall,
                 strategy: 'single',
             },
-            metrics: ctx.metricsEnabled ? finishReport : null,
+            metrics: flags.metricsEnabled ? finishReport : null,
             summary,
         };
-        ctx.stdout.write(`${JSON.stringify(payload, null, 2)}\n`);
-        if (ctx.metricsEnabled && finishReport) {
-            const costUsd = await ctx.estimateCostUsd();
+        io.stdout.write(`${JSON.stringify(payload, null, 2)}\n`);
+        if (flags.metricsEnabled && finishReport) {
+            const costUsd = await hooks.estimateCostUsd();
             writeFinishLine({
-                stderr: ctx.stderr,
-                elapsedMs: Date.now() - ctx.runStartedAtMs,
+                stderr: io.stderr,
+                elapsedMs: Date.now() - flags.runStartedAtMs,
+                elapsedLabel: summaryFromCache ? 'Cached' : null,
                 label: extractionUi.finishSourceLabel,
                 model: usedAttempt.userModelId,
                 report: finishReport,
                 costUsd,
-                detailed: ctx.metricsDetailed,
+                detailed: flags.metricsDetailed,
                 extraParts: buildFinishExtras({
                     extracted,
-                    metricsDetailed: ctx.metricsDetailed,
+                    metricsDetailed: flags.metricsDetailed,
                     transcriptionCostLabel,
                 }),
-                color: ctx.verboseColor,
+                color: flags.verboseColor,
             });
         }
         return;
     }
     if (!summaryAlreadyPrinted) {
-        ctx.clearProgressForStdout();
-        const rendered = !ctx.plain && isRichTty(ctx.stdout)
+        hooks.clearProgressForStdout();
+        const rendered = !flags.plain && isRichTty(io.stdout)
             ? renderMarkdownAnsi(prepareMarkdownForTerminal(summary), {
-                width: markdownRenderWidth(ctx.stdout, ctx.env),
+                width: markdownRenderWidth(io.stdout, io.env),
                 wrap: true,
-                color: supportsColor(ctx.stdout, ctx.envForRun),
+                color: supportsColor(io.stdout, io.envForRun),
                 hyperlinks: true,
             })
             : summary;
-        if (!ctx.plain && isRichTty(ctx.stdout)) {
-            ctx.stdout.write(`\n${rendered.replace(/^\n+/, '')}`);
+        if (!flags.plain && isRichTty(io.stdout)) {
+            io.stdout.write(`\n${rendered.replace(/^\n+/, '')}`);
         }
         else {
-            if (isRichTty(ctx.stdout))
-                ctx.stdout.write('\n');
-            ctx.stdout.write(rendered.replace(/^\n+/, ''));
+            if (isRichTty(io.stdout))
+                io.stdout.write('\n');
+            io.stdout.write(rendered.replace(/^\n+/, ''));
         }
         if (!rendered.endsWith('\n')) {
-            ctx.stdout.write('\n');
+            io.stdout.write('\n');
         }
     }
-    const report = ctx.shouldComputeReport ? await ctx.buildReport() : null;
-    if (ctx.metricsEnabled && report) {
-        const costUsd = await ctx.estimateCostUsd();
+    const report = flags.shouldComputeReport ? await hooks.buildReport() : null;
+    if (flags.metricsEnabled && report) {
+        const costUsd = await hooks.estimateCostUsd();
         writeFinishLine({
-            stderr: ctx.stderr,
-            elapsedMs: Date.now() - ctx.runStartedAtMs,
+            stderr: io.stderr,
+            elapsedMs: Date.now() - flags.runStartedAtMs,
+            elapsedLabel: summaryFromCache ? 'Cached' : null,
             label: extractionUi.finishSourceLabel,
             model: modelMeta.canonical,
             report,
             costUsd,
-            detailed: ctx.metricsDetailed,
+            detailed: flags.metricsDetailed,
             extraParts: buildFinishExtras({
                 extracted,
-                metricsDetailed: ctx.metricsDetailed,
+                metricsDetailed: flags.metricsDetailed,
                 transcriptionCostLabel,
             }),
-            color: ctx.verboseColor,
+            color: flags.verboseColor,
         });
     }
 }