logseq-local-http

name: logseq-local-http description: "Interact with a local Logseq graph through the HTTP API at http://127.0.0.1:12315 by sending authenticated POST /api requests with plugin SDK method names and args. Use when tasks need to search notes, retrieve related/backlinked pages, find journal notes by day or topic, or traverse page/block links in a local Logseq graph. Prefer graph-aware reading: read current page, then linked references from recent to old, then latest journals before concluding."

Logseq Local HTTP

Overview

Use this skill to query and analyze a local Logseq graph through the Logseq HTTP API server.

Quick Start

1. Always run:

python3 scripts/logseq_http.py --help

before any other command to confirm command availability. 2. Confirm Logseq HTTP server is running at http://127.0.0.1:12315/. 3. Get the API token from Logseq settings and export it:

export LOGSEQ_TOKEN="your-token"
export LOGSEQ_BASE_URL="http://127.0.0.1:12315"   # optional

4. Run a quick connectivity check:

python3 scripts/logseq_http.py call --method logseq.Editor.getCurrentPage

5. If you get 401 Unauthorized, generate a new token in Logseq: Open Logseq → Settings → API Access (or Local HTTP API), create/copy a fresh token, then set it:

export LOGSEQ_TOKEN="your-new-token"

Then rerun the helper commands.

API Contract

Use POST /api with these headers:

• Content-Type: application/json
• Authorization: Bearer <token>

Body format:

{
  "method": "logseq.Editor.getPage",
  "args": ["My Page"]
}

Default Reading Strategy

Do not conclude from a single page unless the user explicitly asks for narrow scope.

Use this sequence:

1. Read current page (or requested page) blocks.
2. Read linked references and prioritize recent-to-old.
3. Read recent journal pages to capture latest context shifts.
4. Synthesize across all three sources and then answer.

Core Tasks

Build graph-aware context first

Use this before deeper analysis:

python3 scripts/logseq_http.py context --output compact --linked-limit 40 --journal-limit 7 --journal-days 30

Use explicit page instead of current page:

python3 scripts/logseq_http.py context --page "Project Alpha" --output compact

Search notes

Use the helper for title + block-content search:

python3 scripts/logseq_http.py search --query "project alpha" --output compact --limit 20

How it works:

1. Call logseq.Editor.getAllPages.
2. Match query against page names/titles.
3. Call logseq.Editor.getPageBlocksTree per page and match block content.

Control response size:

• --limit 20 (alias --max-results) caps returned matched pages.
• --offset 20 paginates to the next page of matches.
• --max-block-hits-per-page 3 caps hit blocks per page.
• --output compact|names|full controls payload detail (names keeps page + IDs + counts).
• --format outline prints compact, human-readable output.
• --format jsonl prints one JSON object per line.
• --compact-json prints one-line JSON for piping.

Retrieve related pages

Use backlinks + outgoing links:

python3 scripts/logseq_http.py related --page "Project Alpha"

How it works:

1. Call logseq.Editor.getPageLinkedReferences for backlinks.
2. Call logseq.Editor.getPageBlocksTree and extract [[Page]] links for outgoing links.
3. Optionally call logseq.Editor.getPagesFromNamespace for namespace neighbors.

Find notes from a particular day

Use a date in YYYY-MM-DD:

python3 scripts/logseq_http.py journal --date "2026-03-02"

How it works:

1. Convert date to journal day integer YYYYMMDD.
2. Call logseq.Editor.getAllPages.
3. Filter pages where journal flag is true and journalDay matches.
4. Load page blocks with logseq.Editor.getPageBlocksTree.

Find notes by topic

Use a topic/page/tag name:

python3 scripts/logseq_http.py topic --name "machine-learning"

How it works:

1. Treat topic as a page/tag and call logseq.Editor.getPageLinkedReferences.
2. Call logseq.Editor.getTagObjects and match tag names.

Traverse the note graph

Run bounded BFS from a seed page:

python3 scripts/logseq_http.py traverse --start-page "Project Alpha" --depth 2 --max-nodes 120

How it works:

1. Neighbors come from outgoing [[Page]] links and backlink sources.
2. BFS expands by depth with a visited set and node cap.
3. Returns JSON with nodes and edges.

Method Reference

Read references/logseq-http-recipes.md for method lists and request templates.

Script Reference

Use scripts/logseq_http.py for all API calls and reusable workflows instead of rewriting request code each time.