Choose theme
AI Resources
NanoClaw
NanoClaw is a personal agent system that connects messaging channels to agents running in containers, with workspaces, memory and scheduled tasks.
A Node host routes messages and launches agent sessions. Channel adapters and alternative providers are installed into a user's fork as needed; connecting another chat does not necessarily create another agent or memory store. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
A host for personal agent sessions
The host receives and delivers channel messages while agent containers perform the work. It is an application around model providers, rather than a model checkpoint.
Why it stands out
Channel and agent identity are separate
Chats can feed one conversation, share an agent across separate conversations, or use separate agent groups. Those arrangements share different parts of the workspace and memory.
Availability
A local install with optional account services
The default local image build needs no NanoClaw account. The project also offers an account-based path for its prebuilt image and managed Slack app; model-provider and messaging-service access are separate requirements.
Why it matters
What makes it useful
Choose what a second channel should share with the first. The isolation guide distinguishes a shared conversation from separate conversations using the same agent, and from separate agent groups. Separate conversations still share that agent's workspace and persistent memory. A fact saved from one chat may therefore be recalled in another; separate chat histories alone do not create separate memories.
What to know
Where it fits
For a fork that needs one more channel or another model provider, the README describes skill-installed adapters rather than adding every integration to the core. Channel modules live on the channels branch and alternative providers on providers; /add-<channel> or a provider skill copies the selected modules into the fork. Changing a group's provider setting is separate from installing the provider implementation.
Notable points
What stands out
For a frequent check that usually has nothing to report, the scheduled-task guide offers a Bash script gate before waking the model. Its final output decision wakeAgent: false completes the run without calling the model; true passes the script's data to the agent. Invalid output or a failed script counts as failure, not a quiet successful check. The guide says eight consecutive failures pause the recurring series.
Before using
What to review
For a fork merged with upstream by hand, the 2.4 update notes require the selected credential gateway's installation skill before restarting: setting NANOCLAW_GATEWAY_PROVIDER alone does not install it, and the host refuses to start without a registered gateway. The project's /update-nanoclaw path materializes the selected gateway before cutover.
The same release notes say groups with no explicit model can move to newer defaults, changing cost and usage-limit consumption. They describe an install-wide default and a per-group model choice for retaining a selection. An application update and a model-choice change are separate parts of the upgrade decision.
The WhatsApp adapter has a separate dependency-version issue. NanoClaw issue #3941 reports a Baileys 7.0.0-rc.9 pin; the current channel skill also lists that version. The Baileys publisher's GHSA-qvv5-jq5g-4cgg advisory lists message spoofing and state corruption in affected versions, with 7.0.0-rc12 and 6.7.22 as patched versions. An upstream dependency patch and its adoption in the NanoClaw channel skill are different facts.
Reader fit
Who may find it relevant
For a machine choosing between building locally and fetching the project's prebuilt image, the image guide keeps local builds supported without a NanoClaw account. Fetching the offered image requires a free account. The guide also says it replaces custom Dockerfile edits: additions outside apt and npm cannot be restored through its derived-image package mechanism. It identifies local builds as the path that retains those additions. The account choice and preservation of custom build steps therefore belong in the same decision.
Editorial note
Why LifeHubber lists it
If removing one checkout is the next task, distinguish that checkout's state from shared services. The README says its uninstaller removes that copy's service, containers, app data, logs, agent files and gateway agents, but leaves the shared gateway service, credentials and other NanoClaw copies alone. It offers a dry-run preview and confirmation by group. Removing a checkout is not described as deleting the shared credential store.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
More in AI Agents
Keep browsing this category
Explore more AI agent projects.
Paperclip
paperclipai/paperclip
A self-hosted server and dashboard for coordinating agent teams through companies, goals, roles, issues, heartbeats, budgets, approvals, and persistent activity records.
OpenDots
CopilotKit/OpenDots
An early CopilotKit template for a persistent specialist-agent workspace with editable pages, per-Dot computers, and separately configured conversation, voice, and Slack services.
OpenSeeker
PolarSeeker/OpenSeeker
A search-agent project whose current v2 release provides a 30B checkpoint, evaluation code, and web-search tools, while its retained v1 release includes public training data.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
