Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion website/docs/overview/categories/overview.md
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# Category Overview
# Overview

The Category system is the intelligence core of vLLM Semantic Router, enabling intelligent query classification and routing decisions based on semantic understanding of user inputs.

Expand Down
2 changes: 1 addition & 1 deletion website/docs/overview/semantic-router-overview.md
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# Semantic Router Overview
# What is Semantic Router?

Semantic routers represent a paradigm shift in how we deploy and utilize large language models at scale. By intelligently routing queries to the most appropriate model based on semantic understanding, these systems optimize the critical balance between performance, cost, and quality.

Expand Down
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# Additional Prompt Classification Routing for vLLM Semantic Router
# Prompt Classification Routing

**Related Issues:** [#313](https://github.com/vllm-project/semantic-router/issues/313), [#200](https://github.com/vllm-project/semantic-router/issues/200)

Expand Down Expand Up @@ -212,7 +212,7 @@ The Regex Scanner uses regular expression patterns to detect structured data and
- **Performance:** ~2-5ms for dozens of patterns
- **Use Case:** PII patterns (SSN, credit cards), CVE IDs, email addresses, structured data
- **Safety:** RE2 engine prevents catastrophic backtracking (ReDoS protection)
- **Limitations:** Best for <100 patterns; for larger rule sets, use MCP with Hyperscan
- **Limitations:** Best for fewer than 100 patterns; for larger rule sets, use MCP with Hyperscan

**Example Use:** Detect and block Social Security Numbers, route CVE IDs to security models.

Expand Down Expand Up @@ -347,7 +347,7 @@ The Signal Fusion Layer is the decision-making engine that combines all signals
- **Priority-Based**: Rules evaluated from highest to lowest priority (200 → 0)
- **Short-Circuit**: First matching rule wins, no further evaluation
- **Boolean Expressions**: Combine multiple signal conditions with AND, OR, NOT
- **Flexible Comparisons**: Support ==, !=, >, <, >=, <= for numeric thresholds
- **Flexible Comparisons**: Support `==`, `!=`, `>`, `<`, `>=`, `<=` for numeric thresholds

**Expression Capabilities:**

Expand Down
9 changes: 8 additions & 1 deletion website/sidebars.ts
Original file line number Diff line number Diff line change
Expand Up @@ -54,7 +54,7 @@ const sidebars: SidebarsConfig = {
},
{
type: 'category',
label: 'Tutorials',
label: 'Capacities',
items: [
{
type: 'category',
Expand Down Expand Up @@ -92,6 +92,13 @@ const sidebars: SidebarsConfig = {
},
],
},
{
type: 'category',
label: 'Proposals',
items: [
'proposals/prompt-classification-routing',
],
},
{
type: 'category',
label: 'Model Training',
Expand Down
Loading