Minicod

Minicod is an academic data infrastructure platform designed to make scholarly information easier to access from software applications, research assistants, and AI agents.

The platform provides a unified API over multiple scholarly data sources, currently including Semantic Scholar, OpenAlex, and PubMed. Instead of requiring developers to integrate each provider separately with different APIs, schemas, and authentication approaches, Minicod normalizes supported records into a consistent interface.

Its core functionality is academic paper search. Users can search papers by topic, DOI, natural-language queries, publication year, citation count, journal, open-access availability, and other supported metadata. Search results can include source identifiers and DOI information when provided by the underlying source.

Minicod also provides individual paper lookups, author information, citation relationships, references, and batch operations. Citation-graph functionality allows applications to explore both forward citations and references connected to scholarly publications.

A major feature is the platform's OpenAI-compatible Chat Completions API. Developers can submit a natural-language research request to /v1/chat/completions, and Minicod converts the request into an academic search before returning generated text alongside structured paper records when available.

This makes Minicod suitable as a research-data layer for AI applications. A developer can integrate the API into a research assistant, literature-discovery application, academic chatbot, evidence-search workflow, or other AI system without implementing separate integrations for every scholarly database.

Minicod also provides an MCP server for AI clients. Compatible applications such as Claude Desktop, Cursor, Cline, Windsurf, and Continue can connect Minicod and perform academic searches directly inside an AI conversation.

The MCP integration exposes tools for paper searches, individual paper retrieval, DOI lookup, author information, citation graphs, references, batch paper operations, and other academic-data routes. Each MCP request uses the user's Minicod account and applicable plan limits.

The platform includes a Journal Data capability for browsing available journal-level metadata. Depending on the underlying record, this can include impact factor, JCR quartile, Scimago quartile, CAS quartile, APC information, and predatory-risk signals. Minicod presents only fields available from its supported records rather than claiming universal coverage.

Minicod additionally operates a Skills Marketplace containing reusable research-workflow instruction bundles. Examples include Citation Verifier, Paper Finder, Journal Quality Check, and Lit Review Helper. These skills provide reusable workflow guidance that can be installed or used alongside AI research systems.

The API is designed around standard HTTP access and Bearer-token authentication. Developers can use ordinary HTTP clients without needing a specialized SDK, while the documentation provides REST endpoints and API examples.

Minicod's architecture uses a Go-based backend API server, PostgreSQL for structured data, Redis for caching and rate limiting, and a provider-abstraction layer designed to support multiple scholarly sources.

The platform follows a quota-based pricing model rather than charging separately for individual API calls. Plan quotas reset daily and monthly, while additional one-time call packages can be purchased when more usage is required.