Using the Logseq API and Make.com to hook it and Notion together.
Based on Tools on Tech's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.
Make.com can’t directly access a laptop-hosted Logseq instance because inbound connections are blocked by firewalls and security settings.
Briefing
Lockseq can be wired into Notion through Make.com, but the hard part isn’t the API calls—it’s getting Make.com (running in the cloud) to reach Lockseq (running on a local laptop) without exposing insecure ports. The workflow demonstrated creates a Notion page, then appends multiple lines into a Lockseq page, using a “migrated” checkbox in Notion to control which items get processed.
The central obstacle is network reachability. Make.com can’t access a laptop directly because firewalls and security keep local services off the public internet. A common workaround is ngrok, which tunnels a local HTTP port (in the demo, Lockseq on port 12315) to a public URL so Make can call it. That approach works for testing, but it has drawbacks: the first-time warning screen requires a paid ngrok plan to remove, and the public address changes when the tunnel restarts—forcing updates to Make’s configuration.
For a more stable setup, the demo switches to a VPS-based reverse tunnel. A secure shell (SSH) reverse tunnel connects from the VPS back to the laptop’s Lockseq port, keeping the laptop service reachable only through the encrypted path. The demo emphasizes a key security detail: binding the tunnel to localhost on the laptop side. If the tunnel were bound broadly (e.g., “all interfaces”), it could expose an unencrypted port that makes it easier to obtain authorization keys and abuse the API.
On the VPS, nginx handles HTTPS and routes incoming requests to the tunnel target. The configuration uses a domain (logseqtools on tech.com) with Let’s Encrypt certificates via certbot, and it forces HTTP-to-HTTPS redirection so the Lockseq API never travels unencrypted over the internet. Once this is running, disconnecting the SSH tunnel produces a “bad gateway,” confirming that the API access depends on the secure tunnel.
With connectivity solved, the Make.com portion becomes straightforward API orchestration. The scenario starts by querying a temporary Notion database for rows where migrated is false—useful for debugging because toggling checkboxes controls which records get processed. Make then calls Lockseq’s API to create a page using an authorization token and a page name taken from Notion. After creation, Make reads the returned page identifier and fetches up to 100 lines of content.
Finally, Make appends blocks/lines to the Lockseq page using an “append block in page” endpoint. Each appended line includes a newline character so entries don’t stack incorrectly. The demo loops over the Notion content lines, sending repeated API requests to Lockseq and capturing partial responses so the returned JSON can be inspected during troubleshooting.
The result is a repeatable pipeline: Notion items marked for migration trigger Lockseq page creation and block appending, while the migrated checkbox prevents reruns. The demo also points to broader use cases—using Lockseq as a central idea capture point and distributing structured content across tools like Todoist and Notion—so long as the API access remains secure and stable.
Cornell Notes
The workflow connects a local Logseq instance to cloud automation in Make.com so Notion data can create and update Logseq pages. The main challenge is that Make.com can’t reach a laptop directly due to firewalls, so a tunnel is required. ngrok can work for quick tests, but it changes URLs and shows warnings unless paid. For longer-term use, the demo uses a VPS with nginx for HTTPS plus an SSH reverse tunnel back to the laptop’s Logseq port, with localhost binding to avoid exposing an insecure API. Once connectivity is stable, Make.com queries Notion rows where migrated is false, creates a Logseq page via the API, then appends multiple blocks/lines to that page using the returned page ID.
Why can’t Make.com call Logseq running on a laptop directly, and what are the two tunneling approaches shown?
What security risk does the demo warn about when configuring the SSH tunnel, and how is it mitigated?
How does nginx + Let’s Encrypt fit into the architecture?
How does the Make.com workflow decide which Notion records to process during debugging?
What sequence of Logseq API calls does Make.com perform to build the page content?
Why does the demo mark items as migrated, and what problem does it prevent?
Review Questions
- Compare ngrok and the VPS + SSH reverse tunnel approach in terms of stability, security, and operational overhead.
- In the Make.com flow, where does the Logseq page ID come from, and how is it used in later API calls?
- What role does the migrated checkbox play in preventing duplicate processing, and how does it help debugging?
Key Points
- 1
Make.com can’t directly access a laptop-hosted Logseq instance because inbound connections are blocked by firewalls and security settings.
- 2
ngrok is useful for quick testing but is less suitable for long-term automation due to URL changes and first-time warning screens unless using a paid plan.
- 3
A VPS with nginx for HTTPS plus an SSH reverse tunnel provides stable, encrypted access to the laptop’s Logseq port.
- 4
Binding the SSH tunnel to localhost on the laptop side reduces the risk of exposing an unencrypted or publicly reachable API endpoint.
- 5
Make.com can query Notion for rows where migrated is false to control which items get processed during debugging.
- 6
The automation flow typically creates a Logseq page first, then appends multiple blocks/lines using the returned page identifier.