Remove unused token estimation code (#945)

Author: chrarnoldus

src/app/api/dev/consume-credits/route.ts

@@ -79,8 +79,6 @@ export async function POST(request: NextRequest): Promise<NextResponse> {
       },
       max_tokens: null,
       has_middle_out_transform: null,
-      estimatedInputTokens: 0,
-      estimatedOutputTokens: 0,
       isStreaming: false,
       prior_microdollar_usage: user.microdollars_used,
       posthog_distinct_id: user.google_user_email,

src/app/api/fim/completions/route.ts

@@ -11,7 +11,6 @@ import { getUserFromAuth } from '@/lib/user.server';
 import {
   checkOrganizationModelRestrictions,
   countAndStoreFimUsage,
-  estimateFimTokens,
   extractFimPromptInfo,
   extractFraudAndProjectHeaders,
   invalidRequestResponse,
@@ -122,7 +121,6 @@ export async function POST(request: NextRequest) {
   const taskId = extractHeaderAndLimitLength(request, 'x-kilocode-taskid') ?? undefined;
 
   // Extract properties for usage context
-  const tokenEstimates = estimateFimTokens(requestBody);
   const promptInfo = extractFimPromptInfo(requestBody);
 
   const userByok = organizationId
@@ -136,8 +134,6 @@ export async function POST(request: NextRequest) {
     promptInfo,
     max_tokens: requestBody.max_tokens ?? null,
     has_middle_out_transform: null, // N/A for FIM
-    estimatedInputTokens: tokenEstimates.estimatedInputTokens,
-    estimatedOutputTokens: tokenEstimates.estimatedOutputTokens,
     fraudHeaders,
     isStreaming: requestBody.stream === true,
     organizationId,

src/app/api/openrouter/[...path]/route.ts

@@ -22,7 +22,6 @@ import {
   captureProxyError,
   checkOrganizationModelRestrictions,
   dataCollectionRequiredResponse,
-  estimateChatTokens_ignoringToolDefinitions,
   extractFraudAndProjectHeaders,
   invalidPathResponse,
   invalidRequestResponse,
@@ -268,7 +267,6 @@ export async function POST(request: NextRequest): Promise<NextResponseType<unkno
   }
 
   // Extract properties for usage context
-  const tokenEstimates = estimateChatTokens_ignoringToolDefinitions(requestBodyParsed);
   const promptInfo = extractPromptInfo(requestBodyParsed);
 
   const usageContext: MicrodollarUsageContext = {
@@ -278,8 +276,6 @@ export async function POST(request: NextRequest): Promise<NextResponseType<unkno
     promptInfo,
     max_tokens: requestBodyParsed.max_tokens ?? null,
     has_middle_out_transform: requestBodyParsed.transforms?.includes('middle-out') ?? false,
-    estimatedInputTokens: tokenEstimates.estimatedInputTokens,
-    estimatedOutputTokens: tokenEstimates.estimatedOutputTokens,
     fraudHeaders,
     isStreaming: requestBodyParsed.stream === true,
     organizationId,

src/lib/llm-proxy-helpers.test.ts

@@ -1,9 +1,5 @@
 import { describe, it, expect } from '@jest/globals';
-import {
-  checkOrganizationModelRestrictions,
-  estimateChatTokens_ignoringToolDefinitions,
-} from './llm-proxy-helpers';
-import type { OpenRouterChatCompletionRequest } from './providers/openrouter/types';
+import { checkOrganizationModelRestrictions } from './llm-proxy-helpers';
 
 describe('checkOrganizationModelRestrictions', () => {
   describe('enterprise plan - model deny list restrictions', () => {
@@ -209,58 +205,3 @@ describe('checkOrganizationModelRestrictions', () => {
     });
   });
 });
-
-describe('estimateChatTokens', () => {
-  it('should estimate tokens from valid messages', () => {
-    const body = {
-      model: 'anthropic/claude-3-opus',
-      messages: [
-        { role: 'user', content: 'Hello, how are you?' },
-        { role: 'assistant', content: 'I am doing well, thank you!' },
-      ],
-    } as OpenRouterChatCompletionRequest;
-
-    const result = estimateChatTokens_ignoringToolDefinitions(body);
-
-    expect(result.estimatedInputTokens).toBeGreaterThan(0);
-    expect(result.estimatedOutputTokens).toBeGreaterThan(0);
-  });
-
-  it('should handle missing messages gracefully (regression test for KILOCODE-WEB-5ND)', () => {
-    // This test ensures we don't crash when messages is undefined/null/invalid
-    // which can happen with malformed API requests from abuse attempts
-    const undefinedMessages = { model: 'test' } as OpenRouterChatCompletionRequest;
-    const nullMessages = {
-      model: 'test',
-      messages: null,
-    } as unknown as OpenRouterChatCompletionRequest;
-
-    expect(estimateChatTokens_ignoringToolDefinitions(undefinedMessages)).toEqual({
-      estimatedInputTokens: 0,
-      estimatedOutputTokens: 0,
-    });
-    expect(estimateChatTokens_ignoringToolDefinitions(nullMessages)).toEqual({
-      estimatedInputTokens: 0,
-      estimatedOutputTokens: 0,
-    });
-  });
-
-  it('should handle content parts with undefined text', () => {
-    const body = {
-      model: 'test',
-      messages: [
-        {
-          role: 'user',
-          content: [
-            { type: 'text', text: undefined },
-            { type: 'text', text: 'hello' },
-          ],
-        },
-      ],
-    } as unknown as OpenRouterChatCompletionRequest;
-
-    const result = estimateChatTokens_ignoringToolDefinitions(body);
-    expect(result.estimatedInputTokens).toBeGreaterThan(0);
-    expect(result.estimatedOutputTokens).toBeGreaterThan(0);
-  });
-});

src/lib/llm-proxy-helpers.ts

@@ -353,47 +353,6 @@ export function extractFimPromptInfo(body: { prompt: string; suffix?: string | n
   };
 }
 
-export function estimateChatTokens_ignoringToolDefinitions(body: OpenRouterChatCompletionRequest): {
-  estimatedInputTokens: number;
-  estimatedOutputTokens: number;
-} {
-  if (!body.messages || !Array.isArray(body.messages)) {
-    return { estimatedInputTokens: 0, estimatedOutputTokens: 0 };
-  }
-  const overallLength = body.messages.reduce(
-    (sum, m) =>
-      sum +
-      (typeof m.content === 'string'
-        ? m.content?.length
-        : Array.isArray(m.content)
-          ? m.content
-              .filter(c => c.type === 'text')
-              .map(c => (c.text ?? '').length)
-              .reduce((l, str) => str + 1 + l, 0)
-          : 0),
-    0
-  );
-  return {
-    estimatedInputTokens: overallLength / 4,
-    estimatedOutputTokens: overallLength / 4, // Conservative estimate
-  };
-}
-
-export function estimateFimTokens(body: {
-  prompt: string;
-  suffix?: string | null;
-  max_tokens?: number | null;
-}): {
-  estimatedInputTokens: number;
-  estimatedOutputTokens: number;
-} {
-  const promptLength = body.prompt.length + (body.suffix?.length || 0);
-  return {
-    estimatedInputTokens: promptLength / 4,
-    estimatedOutputTokens: (body.max_tokens || 1024) / 2,
-  };
-}
-
 // ============================================================================
 // FIM-Specific Code
 // ============================================================================

src/lib/processUsage.test.ts

@@ -329,8 +329,6 @@ describe('logMicrodollarUsage', () => {
       },
       max_tokens: 200,
       has_middle_out_transform: true,
-      estimatedInputTokens: 100,
-      estimatedOutputTokens: 100,
       status_code: 200,
       editor_name: null,
       machine_id: null,

src/lib/processUsage.ts

@@ -164,8 +164,6 @@ export type MicrodollarUsageContext = {
   promptInfo: PromptInfo;
   max_tokens: number | null;
   has_middle_out_transform: boolean | null;
-  estimatedInputTokens: number;
-  estimatedOutputTokens: number;
   isStreaming: boolean;
   prior_microdollar_usage: number;
   /** User email for authenticated users - used as PostHog distinctId. Undefined for anonymous users. */

src/tests/helpers/microdollar-usage.helper.ts

@@ -100,8 +100,6 @@ export function createMockUsageContext(
     },
     max_tokens: null,
     has_middle_out_transform: null,
-    estimatedInputTokens: 100,
-    estimatedOutputTokens: 50,
     isStreaming: false,
     prior_microdollar_usage,
     posthog_distinct_id,