# Connect your NanoClaw agent group to Tareno > Source: https://tareno.co/agents/nanoclaw > Language: en > Add Tareno to a NanoClaw agent group over Streamable HTTP, configure credentials through the gateway, and verify access with a read-only account check. NanoClaw runs agents in isolated containers and connects them to your messaging apps. Its current tools documentation supports remote Streamable HTTP MCP servers directly, so Tareno can be registered for the group that needs social tools. Authentication is a separate step: NanoClaw keeps service credentials in its host-side gateway, rather than in the URL or an agent prompt. This guide gives the documented registration path and the checks to perform before a morning briefing saves any drafts. It covers NanoClaw, not NVIDIA’s separate NemoClaw project; a successful authenticated Tareno call from NanoClaw has not yet been recorded. ## Connection facts | Field | Value | | --- | --- | | Product surface | NanoClaw (nanoclaw) | | Available connection | Custom MCP connection | | Connection evidence | documented | | Methods | mcp | | Authentication | client_dependent | | Setup | [Setup guide](https://tareno.co/docs/mcp) | | Install or connect | [Open setup](https://tareno.co/docs/mcp) | | Evidence observed | 2026-10-26 | | Facts reviewed | 2026-10-25 | Publication status and a working connection are separate facts. Evidence for one product surface does not establish installation or compatibility in another surface. ## How do I connect Tareno to NanoClaw? Register https://tareno.co/api/mcp for the intended NanoClaw group with ncl groups config add-mcp-server, using --url for remote Streamable HTTP. Configure the Tareno credential in NanoClaw’s gateway and grant only that group access. Restart the group and ask for list_accounts. The setup is documented; confirm the authenticated call in your installation before saving drafts. ## What is documented, step by step 1. **Know what you are running.** NanoClaw runs each agent in a Linux container with filesystem isolation and Docker as the default runtime. A host process routes messages from your chat apps to the container and delivers replies. Channels such as Telegram are added with skills like add-telegram. 2. **Register the remote MCP server.** Register the hosted Tareno endpoint with the remote --url option for the intended group. Use the group ID from your NanoClaw installation; this config applies to that group, rather than every agent. A bridge is not required for the documented native Streamable HTTP path. 3. **Configure the credential separately.** Configure the Tareno API credential through NanoClaw’s host-side credential gateway using an Authorization Bearer header for tareno.co, then grant access only to the intended group. Use the gateway’s credential setup, not a raw key in --headers, --env, a URL or a prompt. Native remote registration does not itself complete Tareno OAuth. 4. **Test with a read before any draft.** Restart the group so its next container receives the new MCP configuration. Ask it to call `list_accounts` without making changes. An authenticated empty result can mean no social accounts are connected in Tareno; a 401 means authentication still needs attention. Do not start drafts until you can identify the intended profile. Register Tareno for your agent group: ```bash ncl groups config add-mcp-server --id "" --name tareno --url https://tareno.co/api/mcp ``` Apply the configuration to the group: ```bash ncl groups restart --id "" ``` Tareno's remote MCP endpoint: ```json "url": "https://tareno.co/api/mcp" ``` ## Use NanoClaw for a briefing you can act on Once registration, gateway access and a read-only call succeed, the same group can turn approved notes into social drafts. Keep the briefing, your choice of an idea, and the publishing decision as distinct steps. ### A daily briefing that arrives where you chat NanoClaw can run a scheduled task and message you the result in your chat app. A morning briefing of topics for the day, with draft posts for you to react to, fits that shape: you answer in the chat, and drafts go to Tareno only after you say yes. ### Isolation for sensitive material Because each agent runs in its own container with a limited view of the filesystem, you can keep the notes it reads separate from the rest of the machine. That suits a founder who wants a drafting agent to see a folder of approved notes and nothing else. ### A fork you control NanoClaw adds capabilities by copying skill code into your fork. That means the connection to Tareno would be code and config that you can read and review, which is a different trust model from a hosted connector, and it puts the burden of correct setup on you. ## Prompts to copy ### Check the connection ```text Use Tareno to list my connected social accounts with platform and account name. Read only: do not create, schedule or publish anything. If you cannot reach Tareno from this container, say so plainly and tell me which step is missing. ``` ### Morning briefing ```text Every weekday at 07:30, read the notes in the shared folder [notes] and send me three topic ideas for X in this chat, each with one sentence on why now. Do not save anything to Tareno until I reply with the number of the idea I want. ``` ### Draft from my reply ```text I choose idea 2. Write an X post within 280 characters from my notes, show it here and wait. When I say yes, use Tareno to confirm my X account [handle] and save the approved text as a draft, then return the draft ID. ``` ### Read what is waiting ```text Use Tareno to list my drafts and pending actions and tell me what is waiting for me, with the account and the time requested for each one. Do not approve, edit or delete anything, and tell me if the list is empty. ``` ### Request the slot ```text Take the draft [draft ID] and ask Tareno to schedule it for [tomorrow 08:30] in [Europe/Berlin]. Return the action ID and remind me that the request still waits for my approval in the Tareno dashboard. ``` ## Guides by network - X: [What would a NanoClaw morning briefing with X drafts look like through Tareno?](https://tareno.co/agents/nanoclaw/x) ([Markdown](https://tareno.co/agents/nanoclaw/x.md)) ## Good to know before you start - Remote MCP registration is documented by NanoClaw. An authenticated Tareno end-to-end test is still required in your installation; documented setup is not a claim of a successful live connection. - This page covers the container-based NanoClaw. NVIDIA's NemoClaw is a different project with its own MCP documentation, and its steps do not apply here. - Third-party documentation mirrors for NanoClaw exist and may be out of date. Rely on NanoClaw's own repository and docs for syntax. - Keep real Tareno credentials in the host-side gateway and verify that the remote MCP client uses that gateway. Adding --url registers a server; it does not grant access or run an OAuth login. - Scheduling, publishing, edits to scheduled posts and deletions create approval actions in Tareno. The agent can request them but cannot approve them. ## NanoClaw compared with OpenClaw and Hermes Agent OpenClaw and Hermes Agent have Tareno skills published on ClawHub. NanoClaw offers a per-group MCP configuration and a host-side credential gateway. Choose the runtime you already operate, follow its own authentication instructions, and check the actual tools returned by Tareno before starting recurring work. ## Frequently asked questions ### Does NanoClaw support MCP servers? Yes. NanoClaw documents both local stdio servers and remote Streamable HTTP servers. Use ncl groups config add-mcp-server with --id, --name and --url for Tareno. The group receives the server at its next container start, and credentials are configured separately in the gateway. ### Is NanoClaw the same as NemoClaw? No. NanoClaw is the container-based assistant that runs agents through Anthropic's Agent SDK. NemoClaw is NVIDIA's project with sandboxes and its own MCP pages, which accept Streamable HTTP endpoints. Do not mix the two when you follow a guide. ### Has Tareno tested a connection? No authenticated NanoClaw-to-Tareno test has been recorded. The registration syntax is documented, so you can set up the group and its gateway credential, then test list_accounts. Confirm the returned profiles before a task creates any drafts or requests a posting time. ### What must I configure besides the server URL? The Tareno credential and the group’s permission to use it. NanoClaw’s gateway stores the credential on the host and injects authentication into allowed outgoing requests. Verify the gateway, the group assignment and the remote client’s proxy use before expecting authenticated tools. ### Where do NanoClaw skills come from? From your own fork. Skills such as the one for Telegram copy the required modules into it, which means you can read what was added. Other model backends, such as Codex or Ollama, are described as skills too. ### Can NanoClaw publish a post without my approval? Not through Tareno. A scheduling or publishing request becomes an approval action in Tareno, and nothing reaches X until you approve it in the dashboard. The agent can only report that the action is pending and read its status. ### Is a daily briefing safe to automate? Sending yourself ideas is. Saving a draft is a small step, and Tareno still holds any time request until you approve it. Keep the agent's instructions explicit that it must wait for your reply before it saves anything. ### Which Tareno tools would NanoClaw call first? Once a connection exists, the agent would usually call `list_accounts` to confirm the destination, then `create_drafts` to save the text. `schedule_posts` comes last and returns an approval action instead of publishing anything, and `get_action_status` shows what became of it. ## Tareno tools used - `list_accounts` - `create_drafts` - `schedule_posts` - `get_action_status` Approval model: Tareno saves drafts directly. Scheduling, publishing, edits to scheduled posts and deletions become pending actions that a person approves in the Tareno dashboard; no agent can approve them. ## Related - [A daily briefing with X drafts](https://tareno.co/agents/nanoclaw/x) - [Remote MCP setup](https://tareno.co/docs/mcp) - [OpenClaw × Tareno](https://tareno.co/agents/openclaw) - [Hermes Agent × Tareno](https://tareno.co/agents/hermes-agent) - [Paperclip × Tareno](https://tareno.co/agents/paperclip) - [Approval workflows](https://tareno.co/features/approval-workflows)