Repository navigation
feat(chat): read and summarize linked documents via readm3 - #121
Merged
Merged
Conversation
… answering
A link in a chat message ("summarize this https://arxiv.org/pdf/2106.01820")
is read through readm3's reader via /api/chat/read, fitted to the small
models' context (start + end, CHAT_DOC_CHARS, default 7000) and prepended
as this turn's system message, like web search. Only the URL reaches the
server, so E2E chats keep working. READM3_TOKEN (optional) skips readm3's
free hourly allowance.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
ThreatCrush Security Scan40 finding(s) HIGH/CRITICAL: 6 | MEDIUM: 14 | LOW: 20
Snippets are redacted; ThreatCrush never prints matched credential material. |
This was referenced Nov 2, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Makes "can you summarize this for me in both technical and layman terms? https://arxiv.org/pdf/2106.01820\" work in /chat.
lib/doc-urls.js: finds up to 2 http(s) links in the message (trailing punctuation stripped).lib/read-url.js: reads through readm3's new reader (POST https://readm3.com/api/v1/read, feat: document reader: /api/v1/read, hosted MCP read_url, readm3 read (0.32.0) profullstack/readm3#86: PDF/HTML/office/text, public addresses only, free hourly allowance then x402). Fits the text intoCHAT_DOC_CHARS(default 7,000 ≈ 1.8k tokens, so it fits a 4k-context node next to the answer): 75% from the start (title, abstract, intro), 25% from the end (conclusion), with a marker. Builds a system message telling the model to answer from the document and not invent.POST /api/chat/read {url}: 20/hour per IP,CHAT_READ_DAILY_CAP(1000). readm3's 402 becomes a plain "allowance used up" note.chat-view.js: when the message has a link, shows "Reading arxiv.org…", prepends the context for this turn (same path as web search, so E2E chats send only the URL), and lists "Read <title> · PDF, 17 pages" under the answer. Failures show a note and the model answers without it.READM3_TOKEN(already read by the Readm3 export route) makes the reads unmetered.Tests:
tests/chat-read-url.test.js(8). Web-search tests still pass.🤖 Generated with Claude Code