Memorybank

pip install cognis-memorybank
memorybank scan .            # → prioritized findings in seconds

🔎 Example output

Real, reproducible output from the tool — runs offline:

$ memorybank-emit --version
memorybank 0.1.0

$ memorybank-emit --help
usage: memorybank [-h] [--version] [--format {table,json}] [--path PATH]
                  {remember,recall,forget,list,stats} ...

Portable agent memory store.

positional arguments:
  {remember,recall,forget,list,stats}
    remember            store a new memory
    recall              retrieve memories ranked by a query
    forget              delete a memory by id
    list                list every memory
    stats               show bank statistics

options:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --format {table,json}
  --path PATH           path to the JSONL memory bank

$ memorybank-emit stats
{
  "count": 0,
  "halflife_days": 14.0,
  "path": "C:\\Users\\user\\cognis-demo\\memorybank\\memorybank.jsonl",
  "tags": {},
  "total_accesses": 0
}

Blocks above are real memorybank output — reproduce them from a clone.

Usage — step by step

1. Install (Python 3.8+, stdlib only):

   pip install memorybank
   

   The store is a single JSONL file (default memorybank.jsonl, override with --path or MEMORYBANK_PATH).
2. Remember a fact, optionally tagged and weighted:

   memorybank remember "User prefers metric units" --tag prefs --importance 2.0
   

3. Recall the most relevant memories for a query (ranked, recency-aware):

   memorybank recall "what units?" --limit 5 --tag prefs
   

   Add --no-touch to retrieve without updating recency.
4. Read the output — every command emits JSON by default; switch to a human table:

   memorybank --format table list
   memorybank stats                 # bank-wide counts/metrics
   memorybank forget <id>           # delete one memory by id
   

5. Drive it from an agent loop / CI — point each session at its own bank file:

   export MEMORYBANK_PATH=./agent_state/memory.jsonl
   memorybank recall "$TASK" --limit 8 | jq -r '.[].text'
   

   Exits non-zero on error so callers can detect failures.

Contents

• Why memorybank? · Features · Quick start · Example · Architecture · AI stack · How it compares · Integrations · Install anywhere · Related · Contributing

Why memorybank?

agent-memory niche

memorybank is single-purpose, scriptable, and self-hostable: point it at a target, get prioritized results in the format your workflow already speaks (table · JSON · SARIF), gate CI on it, and let agents drive it over MCP.

Features

• ✅ Fast, single-purpose CLI
• ✅ JSON / SARIF output for pipelines
• ✅ CI fail-gate (--fail-on)
• ✅ MCP server for AI agents
• ✅ Runs on Linux/macOS/Windows · Docker · devcontainer
• ✅ Ports in Python, JavaScript, Go, and Rust (ports/)

Quick start

pip install cognis-memorybank
memorybank --version
memorybank scan .                       # scan current project
memorybank scan . --format json         # machine-readable
memorybank scan . --fail-on high        # CI gate (non-zero exit)

Example

$ memorybank scan .
  [HIGH    ] MEM-001  example finding             (./src/app.py)
  [MEDIUM  ] MEM-002  another signal              (./config.yaml)

  2 findings · risk score 5 · 38ms

Architecture

flowchart LR
  IN[MCP server] --> P[memorybank<br/>inspect]
  P --> OUT[findings / policy]

Use it from any AI stack

memorybank is interoperable with every popular way of using AI:

• MCP server — memorybank mcp (Claude Desktop, Cursor, Cognis.Studio, uncensored-fleet)
• OpenAI-compatible / JSON — pipe memorybank scan . --format json into any agent or LLM
• LangChain · CrewAI · AutoGen · LlamaIndex — wrap the CLI/JSON as a tool in one line
• CI / scripts — exit codes + SARIF for non-AI pipelines

How it compares

	Cognis memorybank	agent memory
Self-hostable, no account	✅	varies
Single command, zero config	✅	⚠️
JSON + SARIF for CI	✅	varies
MCP-native (AI agents)	✅	❌
Polyglot ports (JS/Go/Rust)	✅	❌
Open license	✅ COCL	varies

Built in the spirit of agent memory, re-framed the Cognis way. Missing a credit? Open a PR.

Integrations

Pipes into your stack: SARIF for code-scanning, JSON for anything, an MCP server (memorybank mcp) for AI agents, and a webhook forwarder for SIEM/Slack/Jira. See docs/INTEGRATIONS.md.

Install — every way, every platform

pip install "git+https://github.com/cognis-digital/memorybank.git"    # pip (works today)
pipx install "git+https://github.com/cognis-digital/memorybank.git"   # isolated CLI
uv tool install "git+https://github.com/cognis-digital/memorybank.git" # uv
pip install cognis-memorybank                                          # PyPI (when published)
docker run --rm ghcr.io/cognis-digital/memorybank:latest --help        # Docker
brew install cognis-digital/tap/memorybank                             # Homebrew tap
curl -fsSL https://raw.githubusercontent.com/cognis-digital/memorybank/main/install.sh | sh

Linux	macOS	Windows	Docker	Cloud
scripts/setup-linux.sh	scripts/setup-macos.sh	scripts/setup-windows.ps1	docker run ghcr.io/cognis-digital/memorybank	DEPLOY.md (AWS/Azure/GCP/k8s)

Related Cognis tools

• agentsmith — Config-first scaffolding and orchestration for multi-agent workflows
• skillhub — Local skill registry and installer for AI agents
• toolguard — Runtime allowlist and policy for agent tool-calls
• evalbench — Offline LLM / agent eval harness with regression gates
• ragkit — Batteries-included local RAG pipeline — ingest, index, serve
• promptpack — Versioned prompt / template registry with A/B and rollbacks

Explore the suite → 🗂️ all 170+ tools · ⭐ awesome-cognis · 🔗 cognis-sources · 🤖 uncensored-fleet · 🧠 engram

Contributing

PRs, new rules, and demo scenarios are welcome under the collaboration-pull model — see CONTRIBUTING.md and SECURITY.md.

⭐ If memorybank saved you time, star it — it genuinely helps others find it.

Interoperability

{} composes with the 300+ tool Cognis suite — JSON in/out and a shared OpenAI-compatible /v1 backbone. See INTEROP.md for the suite map, composition patterns, and reference stacks.

License

Source-available under the Cognis Open Collaboration License (COCL) v1.0 — free for personal, internal-evaluation, research, and educational use; commercial / production use requires a license ([email protected]). See LICENSE.

---