PaperSeek

Research Question

Data Source

Source Filter Any field

Field/context hint Expand citations Adds forward and backward citation traversal; searches may take longer.

Configuration

LLM, Embedding, and Rerank Service

User session

PaperSeek Service Free daily searches with PaperSeek keys Third-party Service Use authorized provider quota Bring your own Key Use your own provider keys

PaperSeek Service

Daily quota 20 successful searches per day Failures do not consume quota. PaperSeek Service supports OpenAlex, arXiv, Semantic Scholar, PubMed, Google Scholar, Crossref, and computer science top-conference search for registered users.

Auto Automatically choose the best available model. Custom Choose request, embedding, and rerank models.

Auto request model glm-5.2

PaperSeek Service request model PaperSeek Service embedding PaperSeek Service pre-reranking

Free models are provided only for light use by registered individual users. PaperSeek is a personally maintained open-source project, so the free quota is limited and does not provide SLA stability guarantees. For heavy use, use the open-source project to self-deploy and provide your own API key. To request a higher personal account quota, send your name, institution, use case, and UID (shown in the Signed-in account row on the account page opened from the upper-right corner) to the following email:

Verify email to use PaperSeek Service Email-verified accounts get up to 20 successful searches per day in light-use mode. Early users who registered through ModelScope can bind email after sign-in.

Third-party Service

Experimental feature. Calls are charged to your selected third-party provider account quota, and both Hugging Face and ModelScope can also consume your linked custom-key quota.

Third-party provider

ModelScope OpenRouter Hugging Face

Use the selected provider's authorized inference quota. Consumption follows that provider’s policy, including any linked custom key.

Model routing Automatic rotates through available provider models and falls back when needed. Custom model ID Embedding routing Local sparse hashing is the default. External embedding routes are experimental and use the selected provider's quota. Custom embedding model ID Rerank routing Local RRF is the default. External rerank routes are experimental and use the selected provider's quota. Custom reranker model ID

User-authorized model calls PaperSeek uses the selected third-party account authorization for this session. No model API key is stored on the server. Source and citation behavior is configurable in Search settings; OpenAlex key usage follows the configured source key settings. Daily remaining calls: shown after the first ModelScope API Inference response.

Custom model API

Use your own model provider key for this browser session. Anonymous users must also provide their own OpenAlex key for OpenAlex searches.

Session only

Basic configuration

Provider Model API Key

Source OpenAlex Set your own OpenAlex API key below for this source; otherwise the system uses the configured default source keys.

OpenAlex API Key

Get an OpenAlex API key

Open https://openalex.org/
Register or sign in.
Go to API settings
Copy the API Key and paste it here.

Advanced settings Protocol, source keys, and retrieval settings

Model protocol

API Type Base URL

Embedding and reranking

Embedding Provider Embedding Model Embedding Base URL Embedding API Key Reranker Provider Reranker Model Reranker Base URL Reranker API Key

Source settings

OpenAlex Email Crossref Email Semantic Scholar API Key PubMed API Key PubMed Email PubMed Tool Serper API Key WoS API Key WoS DB

System Dashboard

idle