@layer-ai/core 2.0.19 → 2.0.20

package/dist/index.d.ts

@@ -2,6 +2,7 @@ export { default as authRouter } from './routes/v1/auth.js';
 export { default as gatesRouter } from './routes/v1/gates.js';
 export { default as keysRouter } from './routes/v1/keys.js';
 export { default as logsRouter } from './routes/v1/logs.js';
+export { default as chatCompletionsRouter } from './routes/v1/chat-completions.js';
 export { default as completeRouter } from './routes/v2/complete.js';
 export { default as chatRouter } from './routes/v3/chat.js';
 export { default as imageRouter } from './routes/v3/image.js';

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAC9D,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAG5D,OAAO,EAAE,OAAO,IAAI,cAAc,EAAE,MAAM,yBAAyB,CAAC;AAGpE,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAC9D,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAC9D,OAAO,EAAE,OAAO,IAAI,gBAAgB,EAAE,MAAM,2BAA2B,CAAC;AACxE,OAAO,EAAE,OAAO,IAAI,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAC1D,OAAO,EAAE,OAAO,IAAI,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAG1D,OAAO,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AACpD,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAG3C,OAAO,EAAE,EAAE,EAAE,MAAM,sBAAsB,CAAC;AAC1C,OAAO,EAAE,OAAO,IAAI,KAAK,EAAE,MAAM,mBAAmB,CAAC;AAGrD,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,qBAAqB,EAAE,MAAM,qBAAqB,CAAC;AAC9E,YAAY,EAAE,aAAa,EAAE,MAAM,eAAe,CAAC;AAGnD,eAAO,MAAM,gBAAgB,GAAU,QAAQ,MAAM,KAAG,OAAO,CAAC,MAAM,CAGrE,CAAC;AAEF,eAAO,MAAM,wBAAwB,GAAU,QAAQ,MAAM,KAAG,OAAO,CAAC,IAAI,CAG3E,CAAC;AAGF,cAAc,6BAA6B,CAAC;AAG5C,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,QAAQ,EAAE,WAAW,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAC9D,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,qBAAqB,EAAE,MAAM,iCAAiC,CAAC;AAGnF,OAAO,EAAE,OAAO,IAAI,cAAc,EAAE,MAAM,yBAAyB,CAAC;AAGpE,OAAO,EAAE,OAAO,IAAI,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAC9D,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAC9D,OAAO,EAAE,OAAO,IAAI,gBAAgB,EAAE,MAAM,2BAA2B,CAAC;AACxE,OAAO,EAAE,OAAO,IAAI,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAC1D,OAAO,EAAE,OAAO,IAAI,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAG1D,OAAO,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AACpD,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAG3C,OAAO,EAAE,EAAE,EAAE,MAAM,sBAAsB,CAAC;AAC1C,OAAO,EAAE,OAAO,IAAI,KAAK,EAAE,MAAM,mBAAmB,CAAC;AAGrD,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,qBAAqB,EAAE,MAAM,qBAAqB,CAAC;AAC9E,YAAY,EAAE,aAAa,EAAE,MAAM,eAAe,CAAC;AAGnD,eAAO,MAAM,gBAAgB,GAAU,QAAQ,MAAM,KAAG,OAAO,CAAC,MAAM,CAGrE,CAAC;AAEF,eAAO,MAAM,wBAAwB,GAAU,QAAQ,MAAM,KAAG,OAAO,CAAC,IAAI,CAG3E,CAAC;AAGF,cAAc,6BAA6B,CAAC;AAG5C,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,QAAQ,EAAE,WAAW,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC"}

package/dist/index.js

@@ -3,6 +3,7 @@ export { default as authRouter } from './routes/v1/auth.js';
 export { default as gatesRouter } from './routes/v1/gates.js';
 export { default as keysRouter } from './routes/v1/keys.js';
 export { default as logsRouter } from './routes/v1/logs.js';
+export { default as chatCompletionsRouter } from './routes/v1/chat-completions.js';
 // v2 routes
 export { default as completeRouter } from './routes/v2/complete.js';
 // v3 routes

package/dist/lib/openai-conversion.d.ts

@@ -0,0 +1,6 @@
+import type { OpenAIChatCompletionRequest, OpenAIChatCompletionResponse, OpenAIChatCompletionChunk } from '@layer-ai/sdk';
+import type { LayerRequest, LayerResponse } from '@layer-ai/sdk';
+export declare function convertOpenAIRequestToLayer(openaiReq: OpenAIChatCompletionRequest, gateId: string): LayerRequest;
+export declare function convertLayerResponseToOpenAI(layerResp: LayerResponse, requestId?: string): OpenAIChatCompletionResponse;
+export declare function convertLayerChunkToOpenAI(layerChunk: LayerResponse, requestId: string, created: number): OpenAIChatCompletionChunk;
+//# sourceMappingURL=openai-conversion.d.ts.map

package/dist/lib/openai-conversion.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"openai-conversion.d.ts","sourceRoot":"","sources":["../../src/lib/openai-conversion.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAEV,2BAA2B,EAK3B,4BAA4B,EAC5B,yBAAyB,EAE1B,MAAM,eAAe,CAAC;AACvB,OAAO,KAAK,EACV,YAAY,EACZ,aAAa,EAOd,MAAM,eAAe,CAAC;AAqFvB,wBAAgB,2BAA2B,CACzC,SAAS,EAAE,2BAA2B,EACtC,MAAM,EAAE,MAAM,GACb,YAAY,CA2Cd;AAwCD,wBAAgB,4BAA4B,CAC1C,SAAS,EAAE,aAAa,EACxB,SAAS,CAAC,EAAE,MAAM,GACjB,4BAA4B,CA+B9B;AAED,wBAAgB,yBAAyB,CACvC,UAAU,EAAE,aAAa,EACzB,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,GACd,yBAAyB,CA2C3B"}

package/dist/lib/openai-conversion.js

@@ -0,0 +1,215 @@
+import { nanoid } from 'nanoid';
+function convertMessage(openaiMsg) {
+    const layerMsg = {
+        role: openaiMsg.role,
+    };
+    if (typeof openaiMsg.content === 'string') {
+        layerMsg.content = openaiMsg.content;
+    }
+    else if (Array.isArray(openaiMsg.content)) {
+        const textParts = [];
+        const imageParts = [];
+        for (const part of openaiMsg.content) {
+            if (part.type === 'text') {
+                textParts.push(part.text);
+            }
+            else if (part.type === 'image_url') {
+                imageParts.push({
+                    url: part.image_url.url,
+                    detail: part.image_url.detail,
+                });
+            }
+        }
+        if (textParts.length > 0) {
+            layerMsg.content = textParts.join('\n');
+        }
+        if (imageParts.length > 0) {
+            layerMsg.images = imageParts;
+        }
+    }
+    if (openaiMsg.tool_calls) {
+        layerMsg.toolCalls = openaiMsg.tool_calls.map(tc => ({
+            id: tc.id,
+            type: 'function',
+            function: {
+                name: tc.function.name,
+                arguments: tc.function.arguments,
+            },
+        }));
+    }
+    if (openaiMsg.tool_call_id) {
+        layerMsg.toolCallId = openaiMsg.tool_call_id;
+    }
+    if (openaiMsg.name) {
+        layerMsg.name = openaiMsg.name;
+    }
+    return layerMsg;
+}
+function convertTool(openaiTool) {
+    return {
+        type: 'function',
+        function: {
+            name: openaiTool.function.name,
+            description: openaiTool.function.description,
+            parameters: openaiTool.function.parameters,
+        },
+    };
+}
+function convertToolChoice(openaiToolChoice) {
+    if (!openaiToolChoice)
+        return undefined;
+    if (typeof openaiToolChoice === 'string')
+        return openaiToolChoice;
+    return openaiToolChoice;
+}
+function convertResponseFormat(openaiFormat) {
+    if (!openaiFormat)
+        return undefined;
+    if (openaiFormat.type === 'json_schema' && openaiFormat.json_schema) {
+        return {
+            type: 'json_schema',
+            json_schema: openaiFormat.json_schema,
+        };
+    }
+    return openaiFormat.type;
+}
+export function convertOpenAIRequestToLayer(openaiReq, gateId) {
+    let systemPrompt;
+    const messages = [];
+    for (const msg of openaiReq.messages) {
+        if (msg.role === 'system' && typeof msg.content === 'string') {
+            systemPrompt = msg.content;
+        }
+        else {
+            messages.push(convertMessage(msg));
+        }
+    }
+    const layerRequest = {
+        gateId,
+        type: 'chat',
+        model: openaiReq.model,
+        data: {
+            messages,
+            systemPrompt,
+            temperature: openaiReq.temperature,
+            maxTokens: openaiReq.max_tokens || openaiReq.max_completion_tokens,
+            topP: openaiReq.top_p,
+            stream: openaiReq.stream,
+            stopSequences: typeof openaiReq.stop === 'string' ? [openaiReq.stop] : openaiReq.stop,
+            frequencyPenalty: openaiReq.frequency_penalty,
+            presencePenalty: openaiReq.presence_penalty,
+            seed: openaiReq.seed,
+        },
+    };
+    if (openaiReq.tools && openaiReq.tools.length > 0) {
+        layerRequest.data.tools = openaiReq.tools.map(convertTool);
+    }
+    if (openaiReq.tool_choice) {
+        layerRequest.data.toolChoice = convertToolChoice(openaiReq.tool_choice);
+    }
+    if (openaiReq.response_format) {
+        layerRequest.data.responseFormat = convertResponseFormat(openaiReq.response_format);
+    }
+    return layerRequest;
+}
+function convertFinishReason(layerReason) {
+    if (!layerReason)
+        return null;
+    switch (layerReason) {
+        case 'completed':
+            return 'stop';
+        case 'length_limit':
+            return 'length';
+        case 'tool_call':
+            return 'tool_calls';
+        case 'filtered':
+            return 'content_filter';
+        default:
+            return 'stop';
+    }
+}
+function convertToolCallsToOpenAI(layerToolCalls) {
+    if (!layerToolCalls || layerToolCalls.length === 0)
+        return undefined;
+    return layerToolCalls.map(tc => ({
+        id: tc.id,
+        type: 'function',
+        function: {
+            name: tc.function.name,
+            arguments: tc.function.arguments,
+        },
+    }));
+}
+function convertUsage(layerUsage) {
+    return {
+        prompt_tokens: layerUsage?.promptTokens || 0,
+        completion_tokens: layerUsage?.completionTokens || 0,
+        total_tokens: layerUsage?.totalTokens || 0,
+    };
+}
+export function convertLayerResponseToOpenAI(layerResp, requestId) {
+    const id = requestId || layerResp.id || `chatcmpl-${nanoid()}`;
+    const created = layerResp.created || Math.floor(Date.now() / 1000);
+    const message = {
+        role: 'assistant',
+        content: layerResp.content || undefined,
+    };
+    const toolCalls = convertToolCallsToOpenAI(layerResp.toolCalls);
+    if (toolCalls) {
+        message.tool_calls = toolCalls;
+    }
+    const response = {
+        id,
+        object: 'chat.completion',
+        created,
+        model: layerResp.model || 'unknown',
+        choices: [
+            {
+                index: 0,
+                message,
+                finish_reason: convertFinishReason(layerResp.finishReason),
+                logprobs: null,
+            },
+        ],
+        usage: convertUsage(layerResp.usage),
+    };
+    return response;
+}
+export function convertLayerChunkToOpenAI(layerChunk, requestId, created) {
+    const delta = {};
+    if (layerChunk.content && !layerChunk.finishReason) {
+        delta.role = 'assistant';
+    }
+    if (layerChunk.content) {
+        delta.content = layerChunk.content;
+    }
+    if (layerChunk.toolCalls && layerChunk.toolCalls.length > 0) {
+        delta.tool_calls = layerChunk.toolCalls.map((tc, index) => ({
+            index,
+            id: tc.id,
+            type: 'function',
+            function: {
+                name: tc.function.name,
+                arguments: tc.function.arguments,
+            },
+        }));
+    }
+    const chunk = {
+        id: requestId,
+        object: 'chat.completion.chunk',
+        created,
+        model: layerChunk.model || 'unknown',
+        choices: [
+            {
+                index: 0,
+                delta,
+                finish_reason: convertFinishReason(layerChunk.finishReason),
+                logprobs: null,
+            },
+        ],
+    };
+    if (layerChunk.usage && layerChunk.finishReason) {
+        chunk.usage = convertUsage(layerChunk.usage);
+    }
+    return chunk;
+}

package/dist/routes/tests/test-openai-endpoint.d.ts

@@ -0,0 +1,3 @@
+#!/usr/bin/env npx tsx
+export {};
+//# sourceMappingURL=test-openai-endpoint.d.ts.map

package/dist/routes/tests/test-openai-endpoint.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"test-openai-endpoint.d.ts","sourceRoot":"","sources":["../../../src/routes/tests/test-openai-endpoint.ts"],"names":[],"mappings":""}

package/dist/routes/tests/test-openai-endpoint.js

@@ -0,0 +1,292 @@
+#!/usr/bin/env npx tsx
+console.log('='.repeat(80));
+console.log('OPENAI-COMPATIBLE ENDPOINT TESTS');
+console.log('='.repeat(80));
+console.log('');
+const BASE_URL = process.env.API_URL || 'http://localhost:3004';
+const API_KEY = process.env.LAYER_API_KEY;
+const GATE_ID = process.env.TEST_GATE_ID;
+if (!API_KEY) {
+    console.error('❌ Error: LAYER_API_KEY environment variable not set');
+    process.exit(1);
+}
+if (!GATE_ID) {
+    console.error('❌ Error: TEST_GATE_ID environment variable not set');
+    process.exit(1);
+}
+async function testNonStreamingBasic() {
+    console.log('Test 1: Non-streaming basic chat completion');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'gpt-4o',
+        messages: [
+            { role: 'user', content: 'Say "test passed" and nothing else.' }
+        ],
+        max_tokens: 10,
+        gateId: GATE_ID,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    const data = await response.json();
+    console.log('  Response ID:', data.id);
+    console.log('  Model:', data.model);
+    console.log('  Content:', data.choices[0].message.content);
+    console.log('  Finish reason:', data.choices[0].finish_reason);
+    console.log('  Usage:', data.usage);
+    console.log('  ✅ Non-streaming basic test passed\n');
+}
+async function testNonStreamingWithGateIdInHeader() {
+    console.log('Test 2: Non-streaming with gateId in header');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'gpt-4o',
+        messages: [
+            { role: 'user', content: 'Say "header test passed" and nothing else.' }
+        ],
+        max_tokens: 10,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+            'X-Layer-Gate-Id': GATE_ID,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    const data = await response.json();
+    console.log('  Content:', data.choices[0].message.content);
+    console.log('  ✅ Header gateId test passed\n');
+}
+async function testStreamingBasic() {
+    console.log('Test 3: Streaming basic chat completion');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'gpt-4o',
+        messages: [
+            { role: 'user', content: 'Count from 1 to 3, one number per line.' }
+        ],
+        max_tokens: 50,
+        stream: true,
+        gateId: GATE_ID,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    let chunkCount = 0;
+    let fullContent = '';
+    let finalUsage = null;
+    const reader = response.body?.getReader();
+    const decoder = new TextDecoder();
+    if (!reader) {
+        throw new Error('No response body reader');
+    }
+    while (true) {
+        const { done, value } = await reader.read();
+        if (done)
+            break;
+        const text = decoder.decode(value);
+        const lines = text.split('\n').filter(line => line.trim().startsWith('data: '));
+        for (const line of lines) {
+            const data = line.replace('data: ', '').trim();
+            if (data === '[DONE]') {
+                continue;
+            }
+            try {
+                const chunk = JSON.parse(data);
+                chunkCount++;
+                if (chunk.choices[0].delta.content) {
+                    fullContent += chunk.choices[0].delta.content;
+                }
+                if (chunk.usage) {
+                    finalUsage = chunk.usage;
+                }
+            }
+            catch (e) {
+                // Skip invalid JSON
+            }
+        }
+    }
+    console.log('  Chunks received:', chunkCount);
+    console.log('  Full content:', fullContent.trim());
+    console.log('  Final usage:', finalUsage);
+    console.log('  ✅ Streaming basic test passed\n');
+}
+async function testWithToolCalls() {
+    console.log('Test 4: Non-streaming with tool calls');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'gpt-4o',
+        messages: [
+            { role: 'user', content: 'What is the weather in Paris?' }
+        ],
+        tools: [
+            {
+                type: 'function',
+                function: {
+                    name: 'get_weather',
+                    description: 'Get the current weather for a location',
+                    parameters: {
+                        type: 'object',
+                        properties: {
+                            location: {
+                                type: 'string',
+                                description: 'The city and state, e.g. Paris, France',
+                            },
+                        },
+                        required: ['location'],
+                    },
+                },
+            },
+        ],
+        tool_choice: 'auto',
+        gateId: GATE_ID,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    const data = await response.json();
+    console.log('  Finish reason:', data.choices[0].finish_reason);
+    if (data.choices[0].message.tool_calls && data.choices[0].message.tool_calls.length > 0) {
+        console.log('  Tool calls:', JSON.stringify(data.choices[0].message.tool_calls, null, 2));
+        console.log('  ✅ Tool calls test passed\n');
+    }
+    else {
+        console.log('  ⚠️  No tool calls received (model may have chosen not to use tools)\n');
+    }
+}
+async function testClaudeModel() {
+    console.log('Test 5: OpenAI format with Claude model');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'claude-3-7-sonnet-20250219',
+        messages: [
+            { role: 'user', content: 'Say "claude via openai format works" and nothing else.' }
+        ],
+        max_tokens: 20,
+        gateId: GATE_ID,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    const data = await response.json();
+    console.log('  Model:', data.model);
+    console.log('  Content:', data.choices[0].message.content);
+    console.log('  ✅ Claude model test passed\n');
+}
+async function testGeminiModel() {
+    console.log('Test 6: OpenAI format with Gemini model');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'gemini-2.0-flash',
+        messages: [
+            { role: 'user', content: 'Say "gemini via openai format works" and nothing else.' }
+        ],
+        max_tokens: 20,
+        gateId: GATE_ID,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    const data = await response.json();
+    console.log('  Model:', data.model);
+    console.log('  Content:', data.choices[0].message.content);
+    console.log('  ✅ Gemini model test passed\n');
+}
+async function testMistralModel() {
+    console.log('Test 7: OpenAI format with Mistral model');
+    console.log('-'.repeat(80));
+    const request = {
+        model: 'mistral-small-2501',
+        messages: [
+            { role: 'user', content: 'Say "mistral via openai format works" and nothing else.' }
+        ],
+        max_tokens: 20,
+        gateId: GATE_ID,
+    };
+    const response = await fetch(`${BASE_URL}/v1/chat/completions`, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+            'Authorization': `Bearer ${API_KEY}`,
+        },
+        body: JSON.stringify(request),
+    });
+    if (!response.ok) {
+        const error = await response.json();
+        throw new Error(`Request failed: ${JSON.stringify(error)}`);
+    }
+    const data = await response.json();
+    console.log('  Model:', data.model);
+    console.log('  Content:', data.choices[0].message.content);
+    console.log('  ✅ Mistral model test passed\n');
+}
+(async () => {
+    try {
+        await testNonStreamingBasic();
+        await testNonStreamingWithGateIdInHeader();
+        await testStreamingBasic();
+        await testWithToolCalls();
+        await testClaudeModel();
+        await testGeminiModel();
+        await testMistralModel();
+        console.log('='.repeat(80));
+        console.log('✅ ALL OPENAI-COMPATIBLE ENDPOINT TESTS PASSED');
+        console.log('='.repeat(80));
+    }
+    catch (error) {
+        console.error('❌ Test failed:', error);
+        process.exit(1);
+    }
+})();
+export {};

package/dist/routes/v1/chat-completions.d.ts

@@ -0,0 +1,4 @@
+import type { Router as RouterType } from 'express';
+declare const router: RouterType;
+export default router;
+//# sourceMappingURL=chat-completions.d.ts.map

package/dist/routes/v1/chat-completions.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"chat-completions.d.ts","sourceRoot":"","sources":["../../../src/routes/v1/chat-completions.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,MAAM,IAAI,UAAU,EAAE,MAAM,SAAS,CAAC;AAapD,QAAA,MAAM,MAAM,EAAE,UAAqB,CAAC;AA6RpC,eAAe,MAAM,CAAC"}

package/dist/routes/v1/chat-completions.js

@@ -0,0 +1,262 @@
+import { Router } from 'express';
+import { nanoid } from 'nanoid';
+import { db } from '../../lib/db/postgres.js';
+import { authenticate } from '../../middleware/auth.js';
+import { convertOpenAIRequestToLayer, convertLayerResponseToOpenAI, convertLayerChunkToOpenAI, } from '../../lib/openai-conversion.js';
+import { resolveFinalRequest } from '../v3/chat.js';
+import { callAdapter, callAdapterStream } from '../../lib/provider-factory.js';
+const router = Router();
+async function* executeWithRoutingStream(gateConfig, request, userId) {
+    yield* callAdapterStream(request, userId);
+}
+async function executeWithRouting(gateConfig, request, userId) {
+    const result = await callAdapter(request, userId);
+    return { result, modelUsed: request.model };
+}
+router.post('/', authenticate, async (req, res) => {
+    const startTime = Date.now();
+    if (!req.userId) {
+        const error = {
+            error: {
+                message: 'Missing user ID',
+                type: 'authentication_error',
+                code: 'unauthorized',
+            },
+        };
+        res.status(401).json(error);
+        return;
+    }
+    const userId = req.userId;
+    let gateConfig = null;
+    let layerRequest = null;
+    try {
+        const openaiReq = req.body;
+        const gateId = openaiReq.gateId || req.headers['x-layer-gate-id'];
+        if (!gateId) {
+            const error = {
+                error: {
+                    message: 'Missing required field: gateId (provide in request body or X-Layer-Gate-Id header)',
+                    type: 'invalid_request_error',
+                    param: 'gateId',
+                    code: 'missing_field',
+                },
+            };
+            res.status(400).json(error);
+            return;
+        }
+        const isUUID = /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i.test(gateId);
+        if (!isUUID) {
+            const error = {
+                error: {
+                    message: 'gateId must be a valid UUID',
+                    type: 'invalid_request_error',
+                    param: 'gateId',
+                    code: 'invalid_format',
+                },
+            };
+            res.status(400).json(error);
+            return;
+        }
+        gateConfig = await db.getGateByUserAndId(userId, gateId);
+        if (!gateConfig) {
+            const error = {
+                error: {
+                    message: `Gate with ID "${gateId}" not found`,
+                    type: 'invalid_request_error',
+                    param: 'gateId',
+                    code: 'not_found',
+                },
+            };
+            res.status(404).json(error);
+            return;
+        }
+        if (!openaiReq.messages || !Array.isArray(openaiReq.messages) || openaiReq.messages.length === 0) {
+            const error = {
+                error: {
+                    message: 'Missing required field: messages (must be a non-empty array)',
+                    type: 'invalid_request_error',
+                    param: 'messages',
+                    code: 'missing_field',
+                },
+            };
+            res.status(400).json(error);
+            return;
+        }
+        if (gateConfig.taskType && gateConfig.taskType !== 'chat') {
+            console.warn(`[Type Mismatch] Gate "${gateConfig.name}" (${gateConfig.id}) configured for taskType="${gateConfig.taskType}" ` +
+                `but received request to /v1/chat/completions endpoint. Processing as chat request.`);
+        }
+        layerRequest = convertOpenAIRequestToLayer(openaiReq, gateId);
+        const finalRequest = resolveFinalRequest(gateConfig, layerRequest);
+        const isStreaming = finalRequest.data && 'stream' in finalRequest.data && finalRequest.data.stream === true;
+        if (isStreaming) {
+            res.setHeader('Content-Type', 'text/event-stream');
+            res.setHeader('Cache-Control', 'no-cache');
+            res.setHeader('Connection', 'keep-alive');
+            res.setHeader('X-Accel-Buffering', 'no');
+            const requestId = `chatcmpl-${nanoid()}`;
+            const created = Math.floor(Date.now() / 1000);
+            let promptTokens = 0;
+            let completionTokens = 0;
+            let totalCost = 0;
+            let modelUsed = finalRequest.model;
+            try {
+                for await (const layerChunk of executeWithRoutingStream(gateConfig, finalRequest, userId)) {
+                    if (layerChunk.usage) {
+                        promptTokens = layerChunk.usage.promptTokens || 0;
+                        completionTokens = layerChunk.usage.completionTokens || 0;
+                    }
+                    if (layerChunk.cost) {
+                        totalCost = layerChunk.cost;
+                    }
+                    if (layerChunk.model) {
+                        modelUsed = layerChunk.model;
+                    }
+                    const openaiChunk = convertLayerChunkToOpenAI(layerChunk, requestId, created);
+                    res.write(`data: ${JSON.stringify(openaiChunk)}\n\n`);
+                }
+                res.write(`data: [DONE]\n\n`);
+                res.end();
+                const latencyMs = Date.now() - startTime;
+                db.logRequest({
+                    userId,
+                    gateId: gateConfig.id,
+                    gateName: gateConfig.name,
+                    modelRequested: layerRequest.model || gateConfig.model,
+                    modelUsed: modelUsed,
+                    promptTokens,
+                    completionTokens,
+                    totalTokens: promptTokens + completionTokens,
+                    costUsd: totalCost,
+                    latencyMs,
+                    success: true,
+                    errorMessage: null,
+                    userAgent: req.headers['user-agent'] || null,
+                    ipAddress: req.ip || null,
+                    requestPayload: {
+                        gateId: layerRequest.gateId,
+                        type: layerRequest.type,
+                        model: layerRequest.model,
+                        data: layerRequest.data,
+                        metadata: layerRequest.metadata,
+                    },
+                    responsePayload: {
+                        streamed: true,
+                        model: modelUsed,
+                        usage: { promptTokens, completionTokens, totalTokens: promptTokens + completionTokens },
+                        cost: totalCost,
+                    },
+                }).catch(err => console.error('Failed to log request:', err));
+            }
+            catch (streamError) {
+                const errorMessage = streamError instanceof Error ? streamError.message : 'Unknown streaming error';
+                const openaiError = {
+                    error: {
+                        message: errorMessage,
+                        type: 'server_error',
+                        code: 'stream_error',
+                    },
+                };
+                res.write(`data: ${JSON.stringify(openaiError)}\n\n`);
+                res.end();
+                db.logRequest({
+                    userId,
+                    gateId: gateConfig.id,
+                    gateName: gateConfig.name,
+                    modelRequested: layerRequest.model || gateConfig.model,
+                    modelUsed: null,
+                    promptTokens: 0,
+                    completionTokens: 0,
+                    totalTokens: 0,
+                    costUsd: 0,
+                    latencyMs: Date.now() - startTime,
+                    success: false,
+                    errorMessage,
+                    userAgent: req.headers['user-agent'] || null,
+                    ipAddress: req.ip || null,
+                    requestPayload: {
+                        gateId: layerRequest.gateId,
+                        type: layerRequest.type,
+                        model: layerRequest.model,
+                        data: layerRequest.data,
+                        metadata: layerRequest.metadata,
+                    },
+                    responsePayload: null,
+                }).catch(err => console.error('Failed to log request:', err));
+            }
+            return;
+        }
+        const { result, modelUsed } = await executeWithRouting(gateConfig, finalRequest, userId);
+        const latencyMs = Date.now() - startTime;
+        db.logRequest({
+            userId,
+            gateId: gateConfig.id,
+            gateName: gateConfig.name,
+            modelRequested: layerRequest.model || gateConfig.model,
+            modelUsed: modelUsed,
+            promptTokens: result.usage?.promptTokens || 0,
+            completionTokens: result.usage?.completionTokens || 0,
+            totalTokens: result.usage?.totalTokens || 0,
+            costUsd: result.cost || 0,
+            latencyMs,
+            success: true,
+            errorMessage: null,
+            userAgent: req.headers['user-agent'] || null,
+            ipAddress: req.ip || null,
+            requestPayload: {
+                gateId: layerRequest.gateId,
+                type: layerRequest.type,
+                model: layerRequest.model,
+                data: layerRequest.data,
+                metadata: layerRequest.metadata,
+            },
+            responsePayload: {
+                content: result.content,
+                model: result.model,
+                usage: result.usage,
+                cost: result.cost,
+                finishReason: result.finishReason,
+            },
+        }).catch(err => console.error('Failed to log request:', err));
+        const openaiResponse = convertLayerResponseToOpenAI(result);
+        res.json(openaiResponse);
+    }
+    catch (error) {
+        const latencyMs = Date.now() - startTime;
+        const errorMessage = error instanceof Error ? error.message : 'Unknown error';
+        db.logRequest({
+            userId,
+            gateId: gateConfig?.id || null,
+            gateName: gateConfig?.name || null,
+            modelRequested: (layerRequest?.model || gateConfig?.model) || 'unknown',
+            modelUsed: null,
+            promptTokens: 0,
+            completionTokens: 0,
+            totalTokens: 0,
+            costUsd: 0,
+            latencyMs,
+            success: false,
+            errorMessage,
+            userAgent: req.headers['user-agent'] || null,
+            ipAddress: req.ip || null,
+            requestPayload: layerRequest ? {
+                gateId: layerRequest.gateId,
+                type: layerRequest.type,
+                model: layerRequest.model,
+                data: layerRequest.data,
+                metadata: layerRequest.metadata,
+            } : null,
+            responsePayload: null,
+        }).catch(err => console.error('Failed to log request:', err));
+        console.error('OpenAI chat completion error:', error);
+        const openaiError = {
+            error: {
+                message: errorMessage,
+                type: 'server_error',
+                code: 'internal_error',
+            },
+        };
+        res.status(500).json(openaiError);
+    }
+});
+export default router;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@layer-ai/core",
-  "version": "2.0.19",
+  "version": "2.0.20",
   "description": "Core API routes and services for Layer AI",
   "type": "module",
   "main": "./dist/index.js",
@@ -36,7 +36,7 @@
     "nanoid": "^5.0.4",
     "openai": "^4.24.0",
     "pg": "^8.11.3",
-    "@layer-ai/sdk": "^2.5.6"
+    "@layer-ai/sdk": "^2.5.7"
   },
   "devDependencies": {
     "@types/bcryptjs": "^2.4.6",