IndexThinking 0.19.4

IndexThinking

Working Memory Manager for Reasoning-capable LLMs

👁 NuGet
👁 .NET

What It Does

IndexThinking handles the repetitive-but-hard parts of LLM integration:

• Truncation Recovery - Auto-continue when responses hit token limits
• Reasoning Extraction - Unified API for provider-specific thinking formats
• Context Tracking - Session-aware conversation with sliding window
• Token Management - Budget tracking and complexity estimation
• Content Recovery - Repair truncated JSON/code blocks

Scope

IndexThinking manages a single LLM turn, not multi-step workflows.

Quick Start

dotnet add package IndexThinking

// Register services
services.AddIndexThinkingAgents();
services.AddIndexThinkingContext();

// Wrap any IChatClient
var client = new ChatClientBuilder(innerClient)
 .UseIndexThinking()
 .Build(serviceProvider);

// Use normally
var response = await client.GetResponseAsync(messages);

// Access metadata
var thinking = response.GetThinkingContent();
var metrics = response.GetTurnMetrics();

Session-Aware Chat

// Context is automatically tracked and injected
var response = await client.ChatAsync("session-123", "Do that again");

Streaming with Thinking Orchestration

Streaming uses a Collect-and-Yield pattern: chunks are yielded to the caller immediately while buffered internally. After the stream completes, the buffered response is processed through the full orchestration pipeline (reasoning parsing, budget tracking, context tracking).

await foreach (var update in client.GetStreamingResponseAsync(messages))
{
 // Real-time chunks arrive here
 Console.Write(update.Text);

 // The final update contains orchestration metadata
 if (update.AdditionalProperties?.ContainsKey(ThinkingChatClient.TurnResultKey) == true)
 {
 var result = update.AdditionalProperties[ThinkingChatClient.TurnResultKey] as TurnResult;
 Console.WriteLine($"\nTokens: {result?.Metrics.TotalTokens}");
 }
}

Supported Providers

Enabling Reasoning for DeepSeek/vLLM/Qwen

Some providers require explicit reasoning activation:

var options = new ThinkingChatClientOptions
{
 EnableReasoning = true // Adds include_reasoning: true to requests
};

var client = new ChatClientBuilder(innerClient)
 .UseIndexThinking(options)
 .Build(serviceProvider);

Inline Reasoning Stripping (Automatic)

When reasoning is disabled on continuation requests (to avoid double-billing thinking tokens), open-source models sometimes output inline reasoning text instead of properly tagged content. IndexThinking automatically strips these fragments:

• StripLeadingUntaggedReasoning — removes reasoning paragraphs that appear at the start of a continuation fragment (≥2 consecutive reasoning paragraphs required before stripping)
• StripUntaggedReasoning — removes trailing reasoning that appears after the actual answer content (requires ≥200 chars of content before the marker and ≥200 chars of trailing reasoning)

Both methods are called automatically by the continuation pipeline. They are also available as static methods on OpenSourceReasoningParser for manual use.

Documentation

License

MIT License - See for details.