Pro

Preserves Nexus session context across OpenCode compaction events. Injects project/session IDs into the compaction prompt and records each reset as an auditable session entry.

Token usage and cost tracking for Nexus sessions using native OpenCode message data. Auto-appends cost snapshots to session timelines. Zero external dependencies — no Helicone required.

Pre-injection context compression for Nexus MCP tool outputs. Reduces context window token consumption by intercepting large tool responses before they reach the agent.

Classification-driven encryption for documents and project artifacts. Define confidentiality levels per customer or project and enforce encryption policies automatically — from at-rest AES-256 to field-level encryption with external KMS providers.

Transforms uploaded documents (PDF, Markdown, DOCX) into a searchable knowledge graph. An extraction service analyzes structure, entities, and relationships, then stores them in a PostgreSQL-native graph schema — queryable by agents via MCP tools.

Bring your own LLM hardware into the Nexus ecosystem. Connect any vLLM-compatible inference stack and auto-configure it as a dedicated provider in your OpenCode workspace — zero manual config, instant model routing.

Connects Nexus project intelligence to your CI/CD pipeline. Validates PRs against active ADRs and project directives, generates architectural review comments, and reports pipeline results back into Nexus sessions.