Get This Tool
Engram
Summary
Every time your AI coding agent opens a file it has already read twice this session, you are paying for the same tokens again — and if it hit a bug last week, there is nothing stopping it from writing the same bug today.
Engram sits between your IDE and its file reads, maintaining a local SQLite summary of your codebase so agents pull compressed context instead of raw files. The vendor states an 89% measured token reduction. It installs via npm, runs locally with zero cloud dependency, and connects to Claude Code, Cursor, Cline, Continue, Aider, Codex, Windsurf, and Zed through a combination of OpenVSX extensions, an Anthropic plugin, and adapter scripts. The bug-prevention layer surfaces past mistakes from revert history before the agent touches that code path again. This is a passive interceptor, not an agent — it does not plan tasks or run autonomously.
Bottom line: The right call for any team watching token costs climb on a large, stable codebase — less useful when your repo changes so fast that the cached summaries are stale before the next session starts.
Pricing Plans
FreeOpen Source
Full access to engram CLI, local SQLite storage, all IDE integrations, and git-revert mining
- Local-first context indexing
- 89% token reduction on supported projects
- Multi-IDE support
- Mistake pre-mortem warnings
- Apache 2.0 license
View full pricing on github.com →
Pricing may have changed since last verified. Check the official site for current plans.
Community Performance Report Card
No community ratings yet. Be the first to rate this tool!
Community Benchmarks Community
Sign in to submit a benchmarkNo community benchmarks yet. Be the first to share a real-world data point.
Pros
Sign in to edit- Local SQLite storage with no cloud dependency, which means your codebase summary never leaves your machine — relevant for teams under data-residency constraints that rule out cloud-hosted context tools.
- The vendor states an 89% measured token reduction on repeated file reads, so usage-based billing in tools like Cursor or rate-limited Claude Code sessions consume significantly fewer tokens per session.
- Bug-prevention indexing pulls from your repo's revert history, so an agent approaching a previously broken file sees the failure pattern before it writes — instead of repeating it.
- A single context store shared across Claude Code, Cursor, Cline, Continue, Aider, Codex, Windsurf, and Zed, which means switching tools mid-project or running two tools in parallel does not require rebuilding context from scratch.
- Apache 2.0 license with self-hosted operation, so teams can audit the full codebase, fork it, or adapt the adapter layer without negotiating a commercial agreement.
Cons
Sign in to edit- When the codebase changes rapidly — active feature branches, frequent refactors, multiple contributors merging daily — the SQLite summaries drift from the actual file state. The agent works from a compressed snapshot that no longer matches reality. Teams in this situation either rebuild the index on every session (reducing the cost savings) or accept that the context is partially stale.
- The bug-prevention layer depends on revert history existing and being parseable. Greenfield projects or repos with shallow or non-standard Git history get no benefit from that feature — it simply does not fire.
- Engram has no UI, no observability dashboard, and no way to inspect what the agent is actually receiving as context. When an agent produces unexpected output, diagnosing whether the cause is a stale summary requires digging into the SQLite database directly. Teams that need audit trails or explainability for agent decisions will hit this ceiling and move to a tool that exposes its context pipeline.
Community Reviews
Sign in to write a reviewNo reviews yet. Be the first to share your experience.
About
- Platforms
- Node.js (npm); works in Claude Code, Cursor, Cline, Continue, Aider, Codex CLI, Windsurf, Zed
- API Available
- Yes
- Self-Hosted
- Yes
- Last Updated
- 2026-06-01T03:42:37.860Z
Best For
Who it's for
- Teams and individuals using Cursor, Claude Code, or Continue with high token costs
- Developers in projects with large codebases where repeated file reads add up
- Organizations migrating between AI coding tools and wanting unified context
- Projects with existing revert history that need bug-prevention guardrails
What it does well
- Reducing API token costs for AI coding sessions using Claude Code or Cursor
- Preventing repeated bugs by surfacing past mistakes before agents re-commit them
- Maintaining consistent context across long coding sessions without re-reading files
- Cross-IDE context sharing so multiple tools (Claude Code, Cursor, etc.) use the same codebase summary
- Cost optimization in environments with usage-based billing like Cursor or rate-limited Claude Code
Integrations
Discussion Community
Sign in to commentNo discussion yet. Sign in to start the conversation.
Compare Engram
Spotted incorrect or missing data? Join our community of contributors.
Sign Up to ContributeCommunity Notes & Tips Community
Sign in to contributeBe the first to contribute. General notes, observations, gotchas, and tips from people who use this tool day-to-day.
Frequently Asked Questions
- Is Engram free?
- Yes — Engram is fully free to use. There is no paid tier.
- Is Engram open source?
- Yes. Engram is open source — the source repository is at https://github.com/NickCirv/engram.
- Does Engram have an API?
- Yes. Engram exposes a developer API. See the official documentation at https://github.com/nickcirv/engram for details.
- Can I self-host Engram?
- Yes. Engram supports self-hosting on your own infrastructure.
- When was Engram released?
- Engram was first released in 2026.
- What platforms does Engram support?
- Engram is available on: Node.js (npm); works in Claude Code, Cursor, Cline, Continue, Aider, Codex CLI, Windsurf, Zed.
Hours Saved & ROI Stories Community
Sign in to contributeBe the first to contribute. Concrete time/cost savings, with context. e.g. "Cut my code review backlog from 4h to 45m per week."
AI coding tools re-read the same files repeatedly within a session and across sessions, with no memory of what broke last time. Engram addresses this by maintaining a persistent, compressed representation of your codebase in a local SQLite database. When an agent requests file context, Engram intercepts that read and returns a summary instead of the raw content, reducing the token payload the model receives. Revert history feeds a bug-prevention index so that when the agent approaches a file with a known failure pattern, that context surfaces before the agent writes code.
The cross-IDE sharing model is the distinguishing architectural choice. Because the context store is local and IDE-agnostic, the same codebase summary is available to Claude Code, Cursor, Cline, and the rest simultaneously. Teams migrating between tools, or running multiple tools on the same project, do not rebuild context from scratch for each one. The vendor describes this as a ‘context spine’ — a single source of compressed truth that all attached tools draw from.
Engram fits best in projects where the codebase is large, file reads are the dominant token cost, and the repo is stable enough that summaries remain accurate across sessions. It breaks down when the codebase changes rapidly: the SQLite summaries lag behind the actual state, and an agent working from stale context will miss recent refactors. The tool is fully passive — it does not run tasks, loop on tool calls, or take autonomous actions, so it cannot substitute for an agent orchestration layer.
Installation is via npm. IDE integration uses OpenVSX for VS Code-compatible editors and a dedicated plugin for the Anthropic marketplace. The Apache 2.0 license covers all components, and the vendor states there is no cloud component — all data stays local. The GitHub repository shows active issue and pull request traffic, with 131 stars and 12 forks at the time of curation.