Never lose context. Never start over.
Beyond code, it remembers how your architecture evolves and how your engineering unfolds.
English · 简体中文
Make Every Interaction the Starting Point for the Next
Coding agents are good at the task in front of them, but a new session often starts without the architecture, failed attempts, repository rules, or working preferences established before it.
MemoraX Code gives Codex, Claude Code, DeepSeek Harness, and OpenCode a shared memory layer for that context. It can recall prior engineering knowledge, capture reusable lessons from completed work, maintain repository knowledge, and carry your procedures and preferences into future sessions.
The goal is not to remember everything. It is to bring back the small amount of memory relevant to the current task so the agent can reach useful investigation and validation sooner.
Quick Start
Prepare Node.js 20+ (Node.js 24 LTS recommended) and at least one of Codex,
Claude Code, DeepSeek Harness, or OpenCode. Python 3 is required for Repo
Memory operations. Each coding-agent harness retains its own runtime
requirements; current DeepSeek Harness releases require Node.js
^22.19.0 || >=24.0.0. DSH may be installed globally or initialized
beforehand through its official npx workflow.
Install and Connect
1. Install the Package
npm install -g @memorax/memorax-code
2. Run Setup
memorax-code setup
Setup automatically detects supported coding agents and completes an
account-free connection. If you already have a MemoraX account, run
memorax-code setup --existing-account instead. Restart or refresh every
detected coding agent after setup.
Installation Troubleshooting
If the initial setup does not work as expected, check these common cases:
| Symptom | Recommended fix |
|---|---|
| Installation fails with an unsupported Node.js version | Run node --version and upgrade to Node.js 20 or later before reinstalling MemoraX Code. |
| The package installed but setup did not start | This is expected. Run memorax-code setup from a normal interactive terminal. |
| You want to use an existing MemoraX account | Run memorax-code setup --existing-account and enter the requested values locally. Never paste the API key into chats or public issues. |
| Search, retrieval, or writeback is unavailable after setup | Run memorax-code status and memorax-cli status, then follow the detailed troubleshooting guide. |
See Configuration for supported settings and Troubleshooting for detailed diagnostics.
Try Cross-Session Memory
Clone the example repository from the product website, then open Codex, Claude Code, DeepSeek Harness, or OpenCode in the project directory:
git clone https://github.com/SWE-agent/test-repo.git
cd test-repo
Invoke the Skill as $memorax-code in Codex or /memorax-code in Claude Code
or DeepSeek Harness. In OpenCode, ask the agent to use the memorax-code skill
by name. The prompts below use its product name and work in all four clients.
Send these prompts in order in the same session:
- Use the MemoraX Code skill to build Repo Memory, retrieving only the latest 3 issues, pull requests, and commits.
- Review the recent Repo Memory issue: the zeroth number was once calculated incorrectly. Avoid repeating the same problem now.
- Use the MemoraX Code skill to remember the engineering lesson from this coding task.
Close the current conversation, start a new session in the same repository, and send:
Use the MemoraX Code skill to recall the earlier engineering lesson and suggest what to check.
The agent should retrieve the saved lesson and use it to make suggestions for the current repository.
[!TIP] The prompts above are only for quick verification. In normal use, you do not need to invoke the MemoraX Code skill to add memory manually. It writes relevant memory in the background and guides agents to search when useful.
Four Clear Memory Boundaries
| Memory | The question it answers | Examples |
|---|---|---|
| Coding Memory | What engineering lessons should carry into the next task? | Verified fixes, failed approaches, design rationale, pitfalls, and regression checks |
| Repo Memory | What should an agent know about this repository? | Architecture maps, module ownership, entry points, and commit/PR/MR/issue evidence |
| Personal Memory | How should the agent communicate and collaborate with you? | User Profile preferences such as language, tone, explanation depth, and result format |
| Procedure Memory | How should this kind of task be carried out? | Reusable steps, checklists, prerequisites, exceptions, and validation gates |
Personal Memory and Procedure Memory stay in the current repository under
.repo_memory/. When saved content already exists, MemoraX Code compares its
meaning before writing: an equivalent request makes no change; a durable
refinement or conflict updates the matching entry and removes the superseded
wording; an invalid scope is corrected, or the entry is deleted only when it is
wholly obsolete. An explicit forget request deletes only the named preference,
procedure topic, section, or step and leaves unrelated memory unchanged.
One-time task instructions do not change saved memory, and the Agent asks before
writing when the durable intent or target is unclear.
Product Capabilities
| Capability | What it does |
|---|---|
| Background memory writeback | Extracts reusable knowledge from completed turns and writes it to Coding Memory in the background. |
| Preference continuity | Records User Profile preferences and injects them into future tasks on a configured cadence. |
| Procedure reuse | Records reusable task procedures and reminds future agents to apply them. |
| Background Repo Memory maintenance | Automatically organizes repository structure, entry points, and history evidence in the background, then updates them according to policy to reduce repeated searching and summarization. |
| Active memory control | Lets you search and add memory through the bundled MemoraX Code skill or the CLI. |
| Client integration | Integrates with Codex, Claude Code, DeepSeek Harness, and OpenCode to trigger memory retrieval, reminders, and writeback. Automatic quota reminders are currently available in Codex, Claude Code, and OpenCode. |
Your Memory, Your Control
MemoraX is required for cloud-backed memory. Completing setup activates MemoraX search/add and the generated configuration's automatic writeback; there is no second writeback confirmation. Automatic retrieval remains off until explicitly enabled.
Local trace capture is enabled by default for supported clients. Depending on
client capabilities, retained traces under MEMORAX_CODE_HOME may contain
prompts, responses, recalled memory, reminder text, and local paths. Use the
local trace settings to switch to
metadata-only capture or disable a client's trace.
Active memory operations send their query or selected content to MemoraX. Automatic writeback sends selected user instructions and the matching final Agent response from trusted workspace turns for extraction and storage. It does not upload the complete retained client trace artifact or local trace path.
Sign in to MemoraX Console at any time to view, edit, or delete saved memories. MemoraX Cloud does not receive model-provider credentials or local Backend tokens.
Read Configuration for all settings and Security for network, local-data, and retention boundaries.
Update
For a global npm installation:
memorax-code update
The command follows the installed release channel and preserves configuration. Restart or refresh Codex, Claude Code, DeepSeek Harness, and OpenCode when a release changes installed integration assets.
Windows Upgrade Note
If you are upgrading MemoraX Code v0.1.3-v0.1.6 to v0.1.7 on Windows, run:
memorax-code stop
memorax-code update --latest
memorax-code
This one-time step is not required for later upgrades.
Uninstall
Run the product lifecycle before removing the npm package:
memorax-code uninstall
This removes managed integrations and the global package while retaining
MEMORAX_CODE_HOME (default ~/.memorax-code), Claude plugin data, provider
configuration, and memories stored in MemoraX. Remove retained local or cloud
data separately only after reviewing what you still need.
Documentation
Develop and Contribute
Issues and pull requests are welcome. Read CONTRIBUTING.md before making a change, and never include API keys, raw transcripts, private memory, or local trace artifacts in a public report.
License
MemoraX Code is available under the MIT License.