Legal Changelog.
A permanent record of changes to our Terms and Privacy policies.
Version 2.0
+The Enterprise Batch Update — Massive parallelization for high-velocity teams.
- Parallel Batch Execution: Introduced the
batch_executeMCP tool, allowing agents to run up to 10 analysis operations in a single round-trip. This reduces total latency by 80% for complex code maps. - Token-Efficient Structural Outlines: Optimized symbol extraction to support 100+ functions per file with full docstring previews without hitting context limits.
- Topology Performance: Implemented external vs. project dependency grouping, significantly reducing the "visual noise" of stdlib imports in the project graph.
- Refined Review UI: Officially added
dx reviewto the core feature set, providing high-fidelity visual diffs for all team interactions.
Version 1.9
Major search intelligence overhaul focusing on pinpoint accuracy and AI-driven relevance.
- Deep Code Search: We now index code at the functional level. When you search for logic, the engine takes you to the exact line of code instead of just giving you the file.
- Advanced AI Context: Upgraded our core AI model to better handle long, complex code files, providing a significant jump in how the system understands your project's logic.
- Clean Search Filtering: Implemented a smarter filtering layer that automatically hides irrelevant results, saving you from digging through noise.
- Under-the-Hood Stability: Streamlined our internal search pipeline to reduce overhead, making the tool faster and more reliable during heavy use.
Version 1.8
Strategic performance upgrades and enhanced architectural visibility for high-security environments.
- Server-Side Neural Re-ranking: Optimized semantic search by moving AI-powered re-ranking to the server. This reduces client-side CPU load and significantly improves search relevance and precision.
- Granular Synaptic Mapping: Refined AST and Tree-sitter extraction to capture specific imported symbols (Classes, Functions) rather than just module basenames, creating a higher-fidelity project graph.
- High-Security Dependency Tracing: Restored the "Project Graph" visibility for E2EE (Neural Vault) users. The server now performs in-memory decryption and search fallbacks to trace dependencies through encrypted layers securely.
- Enhanced Security Scrubbing: Hardened the
scrub_secretsengine with platform-specific patterns for AWS, GCP, Azure, and GitHub, ensuring credentials never leave the local environment.
Version 1.7
Version 1.6
Version 1.5
Major upgrades to Dalexor MCP Tooling for precise, token-efficient codebase interactions.
- Agentic Neural Search: Re-architected semantic search to securely decrypt results locally and inject exact file line numbers into code summaries.
- Structural Intelligence: Leveraged AST processing to instantly extract exact file skeletons (imports, classes, functions) without loading heavy codebase files.
- Atomic Diffs: Overhauled diff engine to parse logic replays directly from Transaction Memory for real-time debugging accuracy.
- Protocol Stealth: Enforced explicit pure JSON-RPC transport layers to guarantee IDE bridging stability under heavy agent workloads.
Version 1.4
Security hardening and fair-use resource reallocation.
- Refined: API usage tracking is now calculated per-individual user instead of per-organization, ensuring fair resource distribution.
- Added: Source-based credential sharing defense (IP fingerprinting) to prevent unauthorized concurrent session usage.
- Updated: Adjusted Pro tier API rate limits to 150 requests/hour to better match professional developer workflows.
- Fixed: Corrected pricing dashboard logic to display individual subscription tiers strictly, resolving multi-seat plan visibility mirrors.
- Clarified: Updated Pro plan feature set; Neural Vault (E2EE) is now strictly reserved for Expert and Sovereign high-security tiers.
Version 1.3
Decommissioned legacy MCP tools in favor of high-intelligence Neural layers.
- Added:
neural_searchfor semantic-aware codebase discovery. - Added:
trace_logic_heritagefor narrative-based code evolution. - Added:
diff_evolution_snapshotsfor multi-point historical replay. - Added:
get_structural_outlinefor token-efficient file skeleton analysis. - Removed:
find_related_decisions,get_surgical_context, andget_logical_evolution.
Version 1.2
Enhanced recovery tools and real-time team telemetry.
- Added:
dx restorecommand for intent-based code recovery. - Added: Real-time Team Presence pulses in
dx watch. - Added: Project-scoped intelligence broadcasting for collaborative awareness.
- Removed:
dx tailanddx mcpdeprecated in favor of unified bridge architecture.
Version 1.1
Refined encryption transparency and expanded plan accessibility.
- Added: Dedicated "Power User" scaling plan.
- Added: "Sovereign Transition" handling for manual subscription cancellations.
- Clarified: Distinguished Standard Encryption (TLS/AES) for scaling plans from Neural Vault (E2EE) for Expert+ tiers.
- Refined: Updated "PrivateVault" nomenclature to "NeuralVault" across all technical documentation.
Initial Release 1.0
Launch of the Dalexor MI.