What Flowise already keeps
Memory nodes keep a buffer of recent messages per session. They don’t store facts across sessions with history.
Where Auld fits
Create one Custom Tool per endpoint — recall and remember — with a short JavaScript function that calls the REST API, and attach them to the agent node.
| Concern | Flowise | Auld |
|---|---|---|
| Current conversation | Yes | Not needed |
| Facts across sessions and channels | Your code | Yes |
| Which fact is current after a change | Your code | Yes, with history |
| What was true on a date | Your code | Yes (asOf) |
| Who may see a memory | Your code | Identities and audiences |
| Source for each statement | Your code | Returned with every memory |
Implementation sequence
Set variables
Add the base URL and key as Flowise variables.
Create the recall tool
Input schema with
query; function calls/v1/recall.Create the remember tool
Input schema with
fact.Attach to the agent
Describe each tool so the model knows when to use it.
Example
// Flowise Custom Tool: recall_memory (input schema: query: string)
const res = await fetch(`${$vars.AULD_API_BASE}/v1/recall`, {
method: "POST",
headers: { "Authorization": `Bearer ${$vars.AULD_API_KEY}`, "Content-Type": "application/json" },
body: JSON.stringify({ query: $query, viewer: $flow.sessionId, maxTokens: 1500 }),
});
const data = await res.json();
return data.memories.map(m => `${m.text} (${m.validFrom.slice(0, 10)})`).join("\n") || "nothing known";Illustrative example · standard HTTP · check the framework’s current docs for exact signatures
Or connect the MCP server
If your Flowise setup can use MCP tools, point it at the Auld MCP endpoint instead of writing wrappers. The model gets remember, recall, forget and memory_status. Use wrappers when you need to inject the user id yourself or post-process results; use MCP for speed of setup.
Custom integration recipe using standard HTTP/MCP tools
- This recipe wires Flowise’s own tool interface to the Auld REST API or MCP server; there is nothing Auld-specific to install.
- Bind the viewer id from your authentication layer, not from model output.
- Ingestion is asynchronous — poll the job before depending on a just-sent fact in tests.
- Recall returns up to your plan’s token limit (8,000 on Pay as you go and Flex).
Frequently asked questions
Session id as viewer?
Only if sessions map one-to-one to users. Prefer a real user id passed in the flow’s override config.
Error handling?
Return a short message on non-2xx responses so the agent can continue.