Plugins / AI & summaries

Literature Review with LLM

by KaguraTart · github.com/KaguraTart/literature-review-with-LLM

A paper chat workbench that generates Markdown summaries of papers as linked attachments.

Literature Review with LLM has a Zotero Plugins grade of A (recommended) because every code check came out low. It works with Zotero 9, not the current Zotero 10. Last checked 15 Nov 2026.

View source
Downloads a .xpi file (0.4 MB). In Zotero, open Tools → Plugins and install it from the file.
In useAn estimate of how many copies of Zotero have it: Zotero downloads each new version once for every copy that has the plugin, so we count the downloads of its most-downloaded recent release. Rounded. Plugins that update from outside GitHub count low.
660
DownloadsAll-time downloads of its release files from GitHub, counting every update. Installs from elsewhere aren't counted.
1,629
Latest releaseDownloads of version 0.1.146, its latest release, since it came out on 30 Jun 2026. A new release starts from zero; Downloads counts every release.
658
ReleasesVersions published on GitHub. Everyone who has the plugin downloads each one again, which adds up in Downloads.
147
AgeTime since its GitHub repository was created, on 14 Jun 2026. The plugin can be younger than its repository.
5 months
StarsPeople who starred the repository on GitHub: a rough measure of interest.
10
ContributorsPeople who have committed code to this plugin. Bots don't count, nor do the authors of a template or project it was copied from.
2
LicenceThe licence the code is published under, as GitHub reports it.
Apache-2.0
A
Zotero Plugins gradeRecommended
because every code check came out low
Applies to v0.1.146, released 30 Jun 2026Code checked · tested installed in Zotero 9.0.6 on 27 Sep 2026
Doesn't work with the current Zotero (10)Works with Zotero 9.
What we found in the codeSets the gradeThe worst finding in these three areas sets the grade.
Code transparencyReadable code; not yet checked against the published sourceThe developer uploaded the release file from the KaguraTart account. We haven't yet checked whether this file matches the public source code.UnknownWe couldn't check this yet.
Updates
Updates come from this project's GitHub repositoryThe update address offers this version
Where your data goesOnly contacts services we could identify37 addresses in 3 groups.LowLow concern: normal for plugins that do this job. Listed so you know.
AI services (31)
ai-gateway.vercel.shapi.anthropic.comapi.cerebras.aiapi.cline.botapi.cloudflare.comapi.deepinfra.comapi.deepseek.comapi.fireworks.aiapi.groq.comapi.hunyuan.cloud.tencent.comapi.minimaxi.comapi.mistral.aiapi.moonshot.aiapi.openai.comapi.perplexity.aiapi.sambanova.aiapi.siliconflow.cnnamed in its codeThe address is in its code, but we didn't trace a request to it. It may be a link or unused.api.siliconflow.comapi.together.aiapi.x.aiapi.z.aiark.cn-beijing.volces.comdashscope.aliyuncs.comgenerativelanguage.googleapis.comintegrate.api.nvidia.commodels.github.aiopen.bigmodel.cnopenrouter.aiqianfan.baidubce.comqianfan.bj.baidubce.comnamed in its codeThe address is in its code, but we didn't trace a request to it. It may be a link or unused.router.huggingface.co
Scholarly services (4)
api.crossref.orgnamed in its codeThe address is in its code, but we didn't trace a request to it. It may be a link or unused.api.semanticscholar.orgnamed in its codeThe address is in its code, but we didn't trace a request to it. It may be a link or unused.api.unpaywall.orgnamed in its codeThe address is in its code, but we didn't trace a request to it. It may be a link or unused.export.arxiv.orgnamed in its codeThe address is in its code, but we didn't trace a request to it. It may be a link or unused.
On your computer (2)
127.0.0.1localhost
Where we found it
ai-gateway.vercel.shcontent/bootstrap-settings.js · line 60
return { ...common, protocol: "openai_chat", baseURL: "https://ai-gateway.vercel.sh/v1", capabilities: { ...imageCapabilities, pdfBase64: false } };
api.anthropic.comcontent/bootstrap-settings.js · line 51
return { ...common, protocol: "anthropic_messages", baseURL: "https://api.anthropic.com", capabilities: { ...imageCapabilities, pdfBase64: true } };
api.cerebras.aicontent/bootstrap-settings.js · line 159
return { ...common, protocol: "openai_chat", baseURL: "https://api.cerebras.ai/v1", capabilities: commonCapabilities };
Powerful capabilitiesUses the clipboard, and 2 moreLowLow concern: normal for plugins that do this job. Listed so you know.
Low concern (3)
Uses the clipboardStores API keys or passwordsWorks with files on your computer
Before you installShown, not gradedFacts to help you decide. They don't change the grade.
Works withZotero 9 · not 10Doesn't work with the current Zotero (10)MediumMedium concern: worth reading before you install. One medium finding makes the grade B.
Zotero 9 ✓ worksZotero 10 ✗ doesn't workZotero 11 beta not tested yet: Zotero hasn't opened Zotero 11 testing
What you'll needOr: a model on your own computerDetected automatically from its code
No further detail for this area yet.
MaintenanceActive · last release 30 Jun 20262 contributors
No further detail for this area yet.
LanguagesDocumentation in EnglishInterface: English and Chinese
No further detail for this area yet.
We report what we found in the code. Open any area for the evidence.How we gradeReport a problem

What it does

Written by AI from the README

A paper chat workbench that generates Markdown summaries of papers as linked attachments. It answers questions about a selected paper and compares several papers into a literature matrix and collection-level review drafts. Works with many LLM providers, including local ones such as Ollama.

Turns reading and reviewing a set of papers into saved Markdown notes with evidence labels, instead of ad hoc chats. Its developer says it is usable but still early.

What you'll need

Detected automatically
  • Or: a model on your own computerIt can use a local model such as Ollama
  • Zotero 9Not working with Zotero 10

Where your data goes

We found: only contacts services we could identify. 37 addresses in 3 groups.

AI services (31)
ai-gateway.vercel.shapi.anthropic.comapi.cerebras.aiapi.cline.botapi.cloudflare.comapi.deepinfra.comapi.deepseek.comapi.fireworks.aiapi.groq.comapi.hunyuan.cloud.tencent.comapi.minimaxi.comapi.mistral.aiapi.moonshot.aiapi.openai.comapi.perplexity.aiapi.sambanova.aiapi.siliconflow.cnnamed in its codeapi.siliconflow.comapi.together.aiapi.x.aiapi.z.aiark.cn-beijing.volces.comdashscope.aliyuncs.comgenerativelanguage.googleapis.comintegrate.api.nvidia.commodels.github.aiopen.bigmodel.cnopenrouter.aiqianfan.baidubce.comqianfan.bj.baidubce.comnamed in its coderouter.huggingface.co
Scholarly services (4)
api.crossref.orgnamed in its codeapi.semanticscholar.orgnamed in its codeapi.unpaywall.orgnamed in its codeexport.arxiv.orgnamed in its code
On your computer (2)
127.0.0.1localhost
Evidence · file and line
ai-gateway.vercel.shcontent/bootstrap-settings.js · line 60
return { ...common, protocol: "openai_chat", baseURL: "https://ai-gateway.vercel.sh/v1", capabilities: { ...imageCapabilities, pdfBase64: false } };
api.anthropic.comcontent/bootstrap-settings.js · line 51
return { ...common, protocol: "anthropic_messages", baseURL: "https://api.anthropic.com", capabilities: { ...imageCapabilities, pdfBase64: true } };
api.cerebras.aicontent/bootstrap-settings.js · line 159
return { ...common, protocol: "openai_chat", baseURL: "https://api.cerebras.ai/v1", capabilities: commonCapabilities };

Forks and copies

No forks listed.

The developer's response

The developer hasn't responded. Developers can reply here, and their reply sits next to our findings.I'm the developer

Ratings and reviews

What people who use it say. This is separate from the grade, which is what the code does, and never changes it.

No ratings yet.

Write a review
Your rating of Literature Review with LLM v0.1.146

Your review is public, under the name you give, and published with the directory's data under CC BY 4.0. Links are removed. We never store your IP address: a scrambled code made from it is kept for 30 days to catch fake reviews. Every review is checked before it's published; how fake ones are spotted isn't published, so it can't be gamed.

Install Literature Review with LLM
ARecommended
  1. 1Download the plugin fileDownload v0.1.146.xpi · 0.4 MB
  2. 2In Zotero, open Tools › Plugins
  3. 3Click the gear, choose Install Plugin From File… and pick the file
Step-by-step guide with pictures
Or skip the download: use Zotero Atlas

Our free plugin puts this site inside Zotero. Find Literature Review with LLM there and install it in one click; Zotero checks the file against the one we graded.

It never sends your library, notes or files: only your list of plugins, if you say yes. Made by us and graded by the same rules: Zotero Plugins grade B.