Autoloops/greplica

Persistent, searchable engineering memory for AI coding agents. Saves ~50% tokens and ~30% time while planning

What it solves

Coding agents often waste time and tokens re-learning the same repository context—such as architectural decisions, constraints, and failed approaches—across different sessions. Greplica provides a persistent, searchable memory layer that allows agents to query prior knowledge before starting a new task, reducing redundant exploration and improving planning reliability.

How it works

Greplica analyzes repository structures, code, and session transcripts to extract and store durable knowledge as "components," "flows," and "claims."

  1. Knowledge Extraction: It can ingest bundled transcripts from previous agent sessions to bootstrap memory.
  2. Querying: Agents use the greplica graph context command to retrieve a concise Markdown packet containing relevant facts and target files based on a natural language query.
  3. Updating: Hooks or specific commands save new learnings from current sessions back into the memory.
  4. Deployment Modes:
    • Local Mode: Keeps all data on the user's machine using SQLite and local embeddings.
    • Managed Mode: Connects a repository to a shared team memory server, allowing contributors across different forks to share knowledge and reconcile memory updates via GitHub Actions.

Who it’s for

Developers using AI coding agents (such as Claude Code, Codex, Cursor, or OpenHands) who want to increase their agent's efficiency and reduce token costs by maintaining a long-term memory of their codebase.

Highlights

  • Token Efficiency: Benchmarks show token usage reductions of 40-75% on certain planning tasks.
  • Cross-Agent Compatibility: Supports a wide range of platforms including Cursor, Claude Code, and GitHub Copilot CLI.
  • Visual Memory: Includes a greplica graph view command to visualize the current repository memory in a browser.
  • Privacy Options: Offers a fully local mode with no telemetry.
  • Enterprise-Ready: Managed mode supports organization-level access control and OIDC-attested memory reconciliation via GitHub workflows.

Related

  • Project
    rohitg00/agentmemory

    A persistent memory server for AI coding agents that automatically captures and retrieves project context across sessions using hybrid search.

  • Project
    akitaonrails/ai-memory

    ai‑memory is a Rust‑based server that records every observation a coding‑assistant makes (prompts, tool calls, session boundaries) as plain markdown files in a git‑backed wiki. The server can be run locally, in Docker/Podman, or on a homelab box and is designed to be shared across many agents (Claude Code, Codex, Cursor, Gemini CLI, etc.), machines and team members. Hooks posted by agents are consolidated into readable wiki pages at session end; the next session receives a concise handoff so it can continue exactly where the previous one stopped. Storage is human‑readable markdown plus a SQLite index (FTS5, entity and optional vector search). By default it needs no LLM or embedding keys; optional Anthropic/OpenAI integrations add LLM‑generated summaries and semantic search. Multi‑user authentication, audit logs and per‑person attribution are built in. Install via Arch AUR, Docker wrapper, or native binary, start the server, register a client and its hooks, and the system works transparently.

  • Project
    tickernelz/opencode-mem

    A persistent memory system for AI coding agents that uses a local vector database to retain long-term project context and user preferences across sessions.

  • Project
    OthmanAdi/planning-with-files

    A persistent, file-based planning system for AI coding agents that stores task plans and progress on disk to prevent memory loss during context resets.

  • Project
    rohitg00/pro-workflow

    A persistent memory and knowledge plane for AI coding agents that eliminates repetitive corrections through self-correcting SQLite-backed memory and auto-growing research wikis.