# agenticchat

> Turn natural language into executable code — right in your browser. Lightweight AI chat powered by GPT-4o with sandboxed JavaScript execution.

- **URL**: https://www.freshcrate.ai/projects/agenticchat
- **Author**: sauravbhattacharya001
- **Category**: Security
- **Latest version**: `v2.46.0` (2026-05-20)
- **License**: MIT
- **Source**: https://github.com/sauravbhattacharya001/agenticchat
- **Language**: JavaScript
- **GitHub**: 2 stars
- **Registry**: github
- **Tags**: `agentic`, `ai`, `browser`, `chatbot`, `code-execution`, `code-generation`, `developer-tools`, `generative-ai`, `javascript`

## Description

Turn natural language into executable code — right in your browser. Lightweight AI chat powered by GPT-4o with sandboxed JavaScript execution.

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `v2.46.0` | 2026-05-20 | High | # Agentic Chat v2.46.0 — Deadline-tracker perf, SW LRU bound, sandbox hardening, README badge refresh  This release rolls up 10 commits since **v2.45.0** focused on performance, security hardening, test repair, and developer ergonomics.  ## ⚡ Performance - **deadline-tracker:** `_save()` is now only invoked when `_checkUrgent` actually mutated state, eliminating a chatty write loop on every render. (#156, `2179559`) - **service worker:** `_lastRevalidated` is now a bounded **LRU of 512 entries** |
| `v2.44.0` | 2026-05-17 | High | # Agentic Chat v2.44.0 — Action Items, Conversation DNA, Learning Tracker & Cleaner Escapes  A meaty release: three new autonomous "Smart*" modules, a perf pass on the learning tracker, shared HTML-escape consolidation across three more IIFEs, plus code coverage tooling and on-page SEO.  ## ✨ New autonomous modules  - **SmartActionItemExtractor** — surfaces concrete commitments, TODOs, and   next-steps from the live conversation; owner/confidence/kind heuristics so   the panel is actionable inst |
| `v2.43.0` | 2026-05-02 | High | ## What's New  ### ✨ Features  - **SmartIntentAligner** — autonomous intent alignment verification engine with 7 analysis engines (topic drift, scope mismatch, constraint violation, format mismatch, depth mismatch, hallucination risk, relevance decay), composite 0-100 alignment scoring, corrective prompt generation, 4-tab panel, and floating badge. Shortcut: Ctrl+Shift+Y. 53 tests. - **SmartConversationWeather** — weather-metaphor conversation health tracker with 6 atmosphere dimensions, 5 weath |
| `v2.42.0` | 2026-05-01 | High | ## ✨ New Features  - **SmartGoalTracker** — Autonomous goal extraction and progress tracking (Alt+Shift+G) - **SmartPatternAutomator** — Workflow pattern detection and automation engine (Alt+Shift+J) - **SmartCognitiveLoad** — Autonomous cognitive load monitor with overwhelm detection (Alt+Shift+L) - **SmartAdaptiveTone** — Communication style profiler and tone adapter (Alt+Shift+Y)  ## 🔒 Security  - Route SmartFactMemory, SmartPatternAutomator, SmartConversationOracle, SmartGoalTracker, SmartA |
| `v2.40.0` | 2026-04-25 | High | ## What's New  ### ✨ Smart Conversation Digest (Alt+Shift+J) Cross-session summary panel that analyzes all saved sessions and produces structured summaries: - Overview stats (session count, message count, sentiment) - Top topics extracted via TF-IDF tokenization - Action items detected from conversation patterns - Unresolved questions with heuristic answer-matching - Per-session breakdown with sentiment indicators - Markdown export for offline review - Today/7-day/30-day time windows  ### 🌿 Con |
| `v2.32.0` | 2026-04-22 | High | ## What's New  ### ✨ Features - **Smart Conversation Memory** — autonomous cross-session knowledge base that surfaces relevant context from past conversations - **Smart Session Prioritizer (Alt+Shift+P)** — intelligently rank and surface your most important sessions  ### ⚡ Performance - **Spatial grid index for WordCloud** — O(1) collision detection replaces O(placed) linear scan via 64px grid partitioning - **Pre-compiled RegExp in ModelAdvisor** — eliminates up to 39 RegExp allocations per \_a |
| `v2.31.0` | 2026-04-21 | High | ## What's New  ### ✨ Features - **Conversation Momentum Tracker** (Alt+Shift+M) — real-time visualization of conversation pacing and engagement dynamics  ### ⚡ Performance - Cached related-messages and follow-ups in SmartContextSidebar, cached anchor TF-IDF vector and topics in ConversationDriftDetector — eliminates O(M) token-scoring on every MutationObserver fire - Cached entity extraction and stats in SmartContextSidebar render path — avoids O(M×L) regex+concat work during streaming - Single- |
| `v2.30.0` | 2026-04-21 | High | ## What's New  ### ✨ Smart Model Advisor (Alt+Shift+A) Proactive model recommendation engine that analyzes your conversation patterns and suggests optimal model switches based on task complexity, token usage, and response quality.  ### ✨ Smart Context Sidebar (Alt+Shift+I) Proactive intelligence panel that surfaces relevant context, suggestions, and insights alongside your conversations in real time.  ### ⚡ Performance - **Allocation-free sentiment scoring** in MoodRing and ConversationSentiment |
| `v2.29.0` | 2026-04-20 | High | ## What's New  ### ✨ Smart Context Compressor (Alt+Shift+C) Intelligent context window management that compresses conversation history while preserving key information — keeps conversations going longer without losing important context.  ### ⚡ Performance - **Single-loop word counting** in ChatStats and ConversationSummarizer — replaces multi-pass array allocations with a single iteration, reducing GC pressure on long conversations.  **Full Changelog**: https://github.com/sauravbhattacharya001/a |
| `v2.28.4` | 2026-04-20 | High | ## What's Changed  ### Refactor - **Extract SSE line parser and unify streaming/non-streaming API call paths** — Consolidates duplicate SSE parsing logic into a shared parser module, reducing code duplication across streaming and non-streaming call paths. Both paths now share the same robust line-parsing pipeline. (7f4a1c7)  ### Performance - **Eliminate throwaway array allocations in hot paths** — Replaces temporary array allocations (spreads, slices, intermediate maps) in frequently-called fun |

## Dependency audit

- **Score**: 100/100
- **Total deps**: 3
- **Resolved**: 3
- **Unresolved**: 0
- **License conflicts**: 0
- **Warnings**: 0
- **Scanned**: 2026-05-18

## Citation

- HTML: https://www.freshcrate.ai/projects/agenticchat
- Markdown: https://www.freshcrate.ai/projects/agenticchat.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/agenticchat/deps

_Generated by freshcrate.ai. Indexes github releases for AI-agent ecosystem packages._