> ## Documentation Index > Fetch the complete documentation index at: https://getligura.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Save, find, and continue > A practical walkthrough from the first useful result to your next AI session. The examples below use **Elephas Teams Launch**. Replace that name with your own Work. These are requests to a connected AI app, not commands you need to memorize. ## 1. Connect your AI app Open [Ligura](https://app.getligura.com/) and use **Connections** to follow the guide for your AI app. Setup and availability vary by app and account. Review the access you grant. See the guides for [Claude](/docs/connect-claude.md), [ChatGPT](/docs/connect-chatgpt.md), [Grok Bot](/docs/connect-grok-bot.md), and [other apps](/docs/connect-other-apps.md). Ligura does not independently send your saved work to AI models for processing. A connected app retrieves or changes saved work through the access you authorize. Connecting one app does not connect every other app you use. ## 2. Choose the right Work Start with an existing Work when the research, decisions, and next steps belong to the same objective. > Find my Elephas Teams Launch Work in Ligura. If there is more than one possible match, show me the choices before saving anything. If this is a new objective: > Create a Work called Elephas Teams Launch for our team launch planning. A Group is optional. If you use several Workspaces, specify the one you mean as well. ## 3. Save a useful result After completing some useful research or planning, ask the AI to save the result in that Work. > Save this sign-in comparison in Elephas Teams Launch. Include the source links and limitations we discussed. Ligura provides the connected AI with tool descriptions, supported formats, and usage guidance. The AI can choose how to structure the saved content; you don't need to specify an artifact type or its technical fields. Mention a format only when you have a preference. Ask the AI to confirm what was saved and where. You can also open the Work in Ligura to check it. A message saying the AI intends to save something is not the same as a successful save. Keep useful results and context. You don't need to save the entire conversation. ## 4. Finish with enough context to continue When wrapping up a longer session, keep the current situation and unfinished actions clear. > Wrap up Elephas Teams Launch so we can continue next time. Save where we stand, what changed, the unfinished tasks, and suggested next steps. Keep recommendations I haven't approved separate from settled decisions. The AI can use State, Tasks, and a Handoff for this, where available. If Handoffs are unavailable, ask for a short saved wrap-up instead. For a small session, updating the relevant artifact may be enough. ## 5. Resume in a new session In the same or another supported AI app with an authorized Ligura connection: > Continue Elephas Teams Launch. Read the saved context and sign-in comparison. Tell me where we stand before proposing the next step. A Work overview helps the AI orient itself. It still needs to retrieve the relevant saved material before relying on its details. The new session has access to what was saved and permitted, not to an unseen transcript of your earlier chat. If the Work doesn't yet have State or a Handoff, ask the AI to find and read the relevant Pages, Tables, or other artifacts. ## 6. Update the work as it develops When you revise existing research, ask to update the saved item instead of making an unnecessary duplicate. > Update the saved sign-in comparison with this new provider information. Keep it in the same Work. Use the artifact's status to make its purpose clear: | Status | Meaning | | - | - | | Draft | Still being worked on or proposed | | Current | The version intended for use now | | Superseded | Replaced by a newer artifact | | Archived | Kept out of the active set | A task's completion status is separate from these artifact statuses. Finishing a task does not automatically mark an entire plan, decision, or Work complete. ## 7. Turn a repeated process into a Skill If you find yourself explaining the same steps in every session, save them once as a Skill. > Save how we just reviewed this launch as a Ligura skill, so we can follow the same steps next time. Review the Draft in Ligura, mark it Current, and ask for it by name in any connected AI app. See [reuse a procedure with Skills](/docs/skills.md). ## If something is missing | What you notice | What to check | | - | - | | The AI cannot find the Work | Confirm its name, the Workspace, and the connection's access. Ask for possible matches. | | The AI remembers only a summary | Ask it to retrieve the specific artifact before using the details. | | A saved link or file cannot be read | Check access to the original resource and the app's file capabilities. Provide a saved summary if needed. | | An older item doesn't appear in a normal search | Ask to include archived or superseded items when you need them. | | The connection has stopped working | Check Connections in Ligura and follow the app's reconnect guide. | Signing out of Ligura does not revoke connected apps. Use **Connections** to disconnect an app when you want to stop its future access. Copies it already retrieved are governed by that app's own behavior and policies. Continuing with another person? Read [work with your team](/docs/team-collaboration.md) and [sharing and invitation links](/docs/sharing-and-links.md). For help, contact [support@getligura.com](mailto:support@getligura.com). This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.