@mariozechner/pi-ai 0.5.18 → 0.5.20

package/README.md

@@ -28,7 +28,7 @@ import { createLLM } from '@mariozechner/pi-ai';
 
 const llm = createLLM('openai', 'gpt-4o-mini');
 
-const response = await llm.complete({
+const response = await llm.generate({
   messages: [{ role: 'user', content: 'Hello!' }]
 });
 
@@ -48,7 +48,7 @@ import { readFileSync } from 'fs';
 const imageBuffer = readFileSync('image.png');
 const base64Image = imageBuffer.toString('base64');
 
-const response = await llm.complete({
+const response = await llm.generate({
   messages: [{
     role: 'user',
     content: [
@@ -77,7 +77,7 @@ const tools = [{
 const messages = [];
 messages.push({ role: 'user', content: 'What is the weather in Paris?' });
 
-const response = await llm.complete({ messages, tools });
+const response = await llm.generate({ messages, tools });
 messages.push(response);
 
 // Check for tool calls in the content blocks
@@ -99,7 +99,7 @@ for (const call of toolCalls) {
 
 if (toolCalls.length > 0) {
   // Continue conversation with tool results
-  const followUp = await llm.complete({ messages, tools });
+  const followUp = await llm.generate({ messages, tools });
   messages.push(followUp);
 
   // Print text blocks from the response
@@ -114,7 +114,7 @@ if (toolCalls.length > 0) {
 ## Streaming
 
 ```typescript
-const response = await llm.complete({
+const response = await llm.generate({
   messages: [{ role: 'user', content: 'Write a story' }]
 }, {
   onEvent: (event) => {
@@ -157,13 +157,17 @@ const response = await llm.complete({
 
 ## Abort Signal
 
+The abort signal allows you to cancel in-progress requests. When aborted, providers return partial results accumulated up to the cancellation point, including accurate token counts and cost estimates.
+
+### Basic Usage
+
 ```typescript
 const controller = new AbortController();
 
 // Abort after 2 seconds
 setTimeout(() => controller.abort(), 2000);
 
-const response = await llm.complete({
+const response = await llm.generate({
   messages: [{ role: 'user', content: 'Write a long story' }]
 }, {
   signal: controller.signal,
@@ -177,18 +181,132 @@ const response = await llm.complete({
 // Check if the request was aborted
 if (response.stopReason === 'error' && response.error) {
   console.log('Request was aborted:', response.error);
+  console.log('Partial content received:', response.content);
+  console.log('Tokens used:', response.usage);
 } else {
   console.log('Request completed successfully');
 }
 ```
 
+### Partial Results and Token Tracking
+
+When a request is aborted, the API returns an `AssistantMessage` with:
+- `stopReason: 'error'` - Indicates the request was aborted
+- `error: string` - Error message describing the abort
+- `content: array` - **Partial content** accumulated before the abort
+- `usage: object` - **Token counts and costs** (may be incomplete depending on when abort occurred)
+
+```typescript
+// Example: User interrupts a long-running request
+const controller = new AbortController();
+document.getElementById('stop-button').onclick = () => controller.abort();
+
+const response = await llm.generate(context, {
+  signal: controller.signal,
+  onEvent: (e) => {
+    if (e.type === 'text_delta') updateUI(e.delta);
+  }
+});
+
+// Even if aborted, you get:
+// - Partial text that was streamed
+// - Token count (may be partial/estimated)
+// - Cost calculations (may be incomplete)
+console.log(`Generated ${response.content.length} content blocks`);
+console.log(`Estimated ${response.usage.output} output tokens`);
+console.log(`Estimated cost: $${response.usage.cost.total}`);
+```
+
+### Continuing After Abort
+
+Aborted messages can be added to the conversation context and continued in subsequent requests:
+
+```typescript
+const context = {
+  messages: [
+    { role: 'user', content: 'Explain quantum computing in detail' }
+  ]
+};
+
+// First request gets aborted after 2 seconds
+const controller1 = new AbortController();
+setTimeout(() => controller1.abort(), 2000);
+
+const partial = await llm.generate(context, { signal: controller1.signal });
+
+// Add the partial response to context
+context.messages.push(partial);
+context.messages.push({ role: 'user', content: 'Please continue' });
+
+// Continue the conversation
+const continuation = await llm.generate(context);
+```
+
+When an aborted message (with `stopReason: 'error'`) is resubmitted in the context:
+- **OpenAI Responses**: Filters out thinking blocks and tool calls from aborted messages, as API call will fail if incomplete thinking and tool calls are submitted
+- **Anthropic, Google, OpenAI Completions**: Send all blocks as-is (text, thinking, tool calls)
+
+## Cross-Provider Handoffs
+
+The library supports seamless handoffs between different LLM providers within the same conversation. This allows you to switch models mid-conversation while preserving context, including thinking blocks, tool calls, and tool results.
+
+### How It Works
+
+When messages from one provider are sent to a different provider, the library automatically transforms them for compatibility:
+
+- **User and tool result messages** are passed through unchanged
+- **Assistant messages from the same provider/model** are preserved as-is
+- **Assistant messages from different providers** have their thinking blocks converted to text with `<thinking>` tags
+- **Tool calls and regular text** are preserved unchanged
+
+### Example: Multi-Provider Conversation
+
+```typescript
+import { createLLM } from '@mariozechner/pi-ai';
+
+// Start with Claude
+const claude = createLLM('anthropic', 'claude-sonnet-4-0');
+const messages = [];
+
+messages.push({ role: 'user', content: 'What is 25 * 18?' });
+const claudeResponse = await claude.generate({ messages }, {
+  thinking: { enabled: true }
+});
+messages.push(claudeResponse);
+
+// Switch to GPT-5 - it will see Claude's thinking as <thinking> tagged text
+const gpt5 = createLLM('openai', 'gpt-5-mini');
+messages.push({ role: 'user', content: 'Is that calculation correct?' });
+const gptResponse = await gpt5.generate({ messages });
+messages.push(gptResponse);
+
+// Switch to Gemini
+const gemini = createLLM('google', 'gemini-2.5-flash');  
+messages.push({ role: 'user', content: 'What was the original question?' });
+const geminiResponse = await gemini.generate({ messages });
+```
+
+### Provider Compatibility
+
+All providers can handle messages from other providers, including:
+- Text content
+- Tool calls and tool results
+- Thinking/reasoning blocks (transformed to tagged text for cross-provider compatibility)
+- Aborted messages with partial content
+
+This enables flexible workflows where you can:
+- Start with a fast model for initial responses
+- Switch to a more capable model for complex reasoning
+- Use specialized models for specific tasks
+- Maintain conversation continuity across provider outages
+
 ## Provider-Specific Options
 
 ### OpenAI Reasoning (o1, o3)
 ```typescript
 const llm = createLLM('openai', 'o1-mini');
 
-await llm.complete(context, {
+await llm.generate(context, {
   reasoningEffort: 'medium'  // 'minimal' | 'low' | 'medium' | 'high'
 });
 ```
@@ -197,7 +315,7 @@ await llm.complete(context, {
 ```typescript
 const llm = createLLM('anthropic', 'claude-3-5-sonnet-20241022');
 
-await llm.complete(context, {
+await llm.generate(context, {
   thinking: {
     enabled: true,
     budgetTokens: 2048  // Optional thinking token limit
@@ -209,7 +327,7 @@ await llm.complete(context, {
 ```typescript
 const llm = createLLM('google', 'gemini-2.5-pro');
 
-await llm.complete(context, {
+await llm.generate(context, {
   thinking: { enabled: true }
 });
 ```

package/dist/models.generated.d.ts

@@ -1593,7 +1593,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "meta-llama/llama-3.1-70b-instruct": {
+            readonly "meta-llama/llama-3.1-405b-instruct": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1609,7 +1609,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "meta-llama/llama-3.1-405b-instruct": {
+            readonly "meta-llama/llama-3.1-70b-instruct": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1641,7 +1641,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "mistralai/mistral-7b-instruct-v0.3": {
+            readonly "mistralai/mistral-7b-instruct:free": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1657,7 +1657,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "mistralai/mistral-7b-instruct:free": {
+            readonly "mistralai/mistral-7b-instruct": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1673,7 +1673,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "mistralai/mistral-7b-instruct": {
+            readonly "mistralai/mistral-7b-instruct-v0.3": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1721,7 +1721,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "meta-llama/llama-3-70b-instruct": {
+            readonly "meta-llama/llama-3-8b-instruct": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1737,7 +1737,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "meta-llama/llama-3-8b-instruct": {
+            readonly "meta-llama/llama-3-70b-instruct": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1849,7 +1849,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "mistralai/mistral-tiny": {
+            readonly "mistralai/mistral-small": {
                 id: string;
                 name: string;
                 provider: string;
@@ -1865,7 +1865,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "mistralai/mistral-small": {
+            readonly "mistralai/mistral-tiny": {
                 id: string;
                 name: string;
                 provider: string;
@@ -2468,7 +2468,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "gpt-4": {
+            readonly "gpt-3.5-turbo": {
                 id: string;
                 name: string;
                 provider: string;
@@ -2483,7 +2483,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "gpt-4-0314": {
+            readonly "gpt-4": {
                 id: string;
                 name: string;
                 provider: string;
@@ -2498,7 +2498,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "gpt-3.5-turbo": {
+            readonly "gpt-4-0314": {
                 id: string;
                 name: string;
                 provider: string;
@@ -2592,7 +2592,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "claude-3-5-haiku-20241022": {
+            readonly "claude-3-5-haiku-latest": {
                 id: string;
                 name: string;
                 provider: string;
@@ -2607,7 +2607,7 @@ export declare const PROVIDERS: {
                 contextWindow: number;
                 maxTokens: number;
             };
-            readonly "claude-3-5-haiku-latest": {
+            readonly "claude-3-5-haiku-20241022": {
                 id: string;
                 name: string;
                 provider: string;

package/dist/models.generated.js

@@ -1595,36 +1595,36 @@ export const PROVIDERS = {
                 contextWindow: 131072,
                 maxTokens: 16384,
             },
-            "meta-llama/llama-3.1-70b-instruct": {
-                id: "meta-llama/llama-3.1-70b-instruct",
-                name: "Meta: Llama 3.1 70B Instruct",
+            "meta-llama/llama-3.1-405b-instruct": {
+                id: "meta-llama/llama-3.1-405b-instruct",
+                name: "Meta: Llama 3.1 405B Instruct",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.09999999999999999,
-                    output: 0.28,
+                    input: 0.7999999999999999,
+                    output: 0.7999999999999999,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
-                contextWindow: 131072,
+                contextWindow: 32768,
                 maxTokens: 16384,
             },
-            "meta-llama/llama-3.1-405b-instruct": {
-                id: "meta-llama/llama-3.1-405b-instruct",
-                name: "Meta: Llama 3.1 405B Instruct",
+            "meta-llama/llama-3.1-70b-instruct": {
+                id: "meta-llama/llama-3.1-70b-instruct",
+                name: "Meta: Llama 3.1 70B Instruct",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.7999999999999999,
-                    output: 0.7999999999999999,
+                    input: 0.09999999999999999,
+                    output: 0.28,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
-                contextWindow: 32768,
+                contextWindow: 131072,
                 maxTokens: 16384,
             },
             "mistralai/mistral-nemo": {
@@ -1643,41 +1643,41 @@ export const PROVIDERS = {
                 contextWindow: 32000,
                 maxTokens: 4096,
             },
-            "mistralai/mistral-7b-instruct-v0.3": {
-                id: "mistralai/mistral-7b-instruct-v0.3",
-                name: "Mistral: Mistral 7B Instruct v0.3",
+            "mistralai/mistral-7b-instruct:free": {
+                id: "mistralai/mistral-7b-instruct:free",
+                name: "Mistral: Mistral 7B Instruct (free)",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.028,
-                    output: 0.054,
+                    input: 0,
+                    output: 0,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
                 contextWindow: 32768,
                 maxTokens: 16384,
             },
-            "mistralai/mistral-7b-instruct:free": {
-                id: "mistralai/mistral-7b-instruct:free",
-                name: "Mistral: Mistral 7B Instruct (free)",
+            "mistralai/mistral-7b-instruct": {
+                id: "mistralai/mistral-7b-instruct",
+                name: "Mistral: Mistral 7B Instruct",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0,
-                    output: 0,
+                    input: 0.028,
+                    output: 0.054,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
                 contextWindow: 32768,
                 maxTokens: 16384,
             },
-            "mistralai/mistral-7b-instruct": {
-                id: "mistralai/mistral-7b-instruct",
-                name: "Mistral: Mistral 7B Instruct",
+            "mistralai/mistral-7b-instruct-v0.3": {
+                id: "mistralai/mistral-7b-instruct-v0.3",
+                name: "Mistral: Mistral 7B Instruct v0.3",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
@@ -1723,32 +1723,32 @@ export const PROVIDERS = {
                 contextWindow: 128000,
                 maxTokens: 4096,
             },
-            "meta-llama/llama-3-70b-instruct": {
-                id: "meta-llama/llama-3-70b-instruct",
-                name: "Meta: Llama 3 70B Instruct",
+            "meta-llama/llama-3-8b-instruct": {
+                id: "meta-llama/llama-3-8b-instruct",
+                name: "Meta: Llama 3 8B Instruct",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.3,
-                    output: 0.39999999999999997,
+                    input: 0.03,
+                    output: 0.06,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
                 contextWindow: 8192,
                 maxTokens: 16384,
             },
-            "meta-llama/llama-3-8b-instruct": {
-                id: "meta-llama/llama-3-8b-instruct",
-                name: "Meta: Llama 3 8B Instruct",
+            "meta-llama/llama-3-70b-instruct": {
+                id: "meta-llama/llama-3-70b-instruct",
+                name: "Meta: Llama 3 70B Instruct",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.03,
-                    output: 0.06,
+                    input: 0.3,
+                    output: 0.39999999999999997,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
@@ -1851,32 +1851,32 @@ export const PROVIDERS = {
                 contextWindow: 128000,
                 maxTokens: 4096,
             },
-            "mistralai/mistral-tiny": {
-                id: "mistralai/mistral-tiny",
-                name: "Mistral Tiny",
+            "mistralai/mistral-small": {
+                id: "mistralai/mistral-small",
+                name: "Mistral Small",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.25,
-                    output: 0.25,
+                    input: 0.19999999999999998,
+                    output: 0.6,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
                 contextWindow: 32768,
                 maxTokens: 4096,
             },
-            "mistralai/mistral-small": {
-                id: "mistralai/mistral-small",
-                name: "Mistral Small",
+            "mistralai/mistral-tiny": {
+                id: "mistralai/mistral-tiny",
+                name: "Mistral Tiny",
                 provider: "openrouter",
                 baseUrl: "https://openrouter.ai/api/v1",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.19999999999999998,
-                    output: 0.6,
+                    input: 0.25,
+                    output: 0.25,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
@@ -2470,24 +2470,24 @@ export const PROVIDERS = {
                 contextWindow: 16385,
                 maxTokens: 4096,
             },
-            "gpt-4": {
-                id: "gpt-4",
-                name: "OpenAI: GPT-4",
+            "gpt-3.5-turbo": {
+                id: "gpt-3.5-turbo",
+                name: "OpenAI: GPT-3.5 Turbo",
                 provider: "openai",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 30,
-                    output: 60,
+                    input: 0.5,
+                    output: 1.5,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
-                contextWindow: 8191,
+                contextWindow: 16385,
                 maxTokens: 4096,
             },
-            "gpt-4-0314": {
-                id: "gpt-4-0314",
-                name: "OpenAI: GPT-4 (older v0314)",
+            "gpt-4": {
+                id: "gpt-4",
+                name: "OpenAI: GPT-4",
                 provider: "openai",
                 reasoning: false,
                 input: ["text"],
@@ -2500,19 +2500,19 @@ export const PROVIDERS = {
                 contextWindow: 8191,
                 maxTokens: 4096,
             },
-            "gpt-3.5-turbo": {
-                id: "gpt-3.5-turbo",
-                name: "OpenAI: GPT-3.5 Turbo",
+            "gpt-4-0314": {
+                id: "gpt-4-0314",
+                name: "OpenAI: GPT-4 (older v0314)",
                 provider: "openai",
                 reasoning: false,
                 input: ["text"],
                 cost: {
-                    input: 0.5,
-                    output: 1.5,
+                    input: 30,
+                    output: 60,
                     cacheRead: 0,
                     cacheWrite: 0,
                 },
-                contextWindow: 16385,
+                contextWindow: 8191,
                 maxTokens: 4096,
             },
         }
@@ -2594,9 +2594,9 @@ export const PROVIDERS = {
                 contextWindow: 200000,
                 maxTokens: 64000,
             },
-            "claude-3-5-haiku-20241022": {
-                id: "claude-3-5-haiku-20241022",
-                name: "Anthropic: Claude 3.5 Haiku (2024-10-22)",
+            "claude-3-5-haiku-latest": {
+                id: "claude-3-5-haiku-latest",
+                name: "Anthropic: Claude 3.5 Haiku",
                 provider: "anthropic",
                 reasoning: false,
                 input: ["text", "image"],
@@ -2609,9 +2609,9 @@ export const PROVIDERS = {
                 contextWindow: 200000,
                 maxTokens: 8192,
             },
-            "claude-3-5-haiku-latest": {
-                id: "claude-3-5-haiku-latest",
-                name: "Anthropic: Claude 3.5 Haiku",
+            "claude-3-5-haiku-20241022": {
+                id: "claude-3-5-haiku-20241022",
+                name: "Anthropic: Claude 3.5 Haiku (2024-10-22)",
                 provider: "anthropic",
                 reasoning: false,
                 input: ["text", "image"],