# zaidmukaddam/scira

Scira (Formerly MiniPerplx) is a minimalistic AI-powered search engine that helps you find information on the internet and cites it too. Powered by Vercel AI SDK!

Repository: https://github.com/zaidmukaddam/scira
Canonical: https://ross.abutalabs.com/products/scira
Homepage: https://scira.ai
Language: TypeScript
License: AGPL-3.0
License Family: copyleft
Topics: ai-search-engine, scira, search-engine, minimalistic-ai-search-engine, scira-ai
Last push: 2026-08-12T22:28:59+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 97, release rhythm 35, longevity 54
- inputs: {"age_days": 756, "days_push": 21, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 11850, forks 1490 (observed 2026-08-28T04:10:50.275564+00:00)

## What it is
Scira is an open-source, minimalistic AI-powered search engine and agentic research platform that plans multi-step searches, retrieves live web sources, and returns answers with inline citations. It offers 17+ search modes (web, academic, X, Reddit, stocks, etc.), scheduled research agents, and MCP-based app integrations, built on the Vercel AI SDK.

## Use cases
- search the web with an AI that cites its sources
- run deep research on a complex question with multiple sources
- search reddit and X posts for community opinions
- find academic papers and scholarly citations
- monitor a topic daily and get emailed updates
- self-host an open-source Perplexity alternative
- fact-check claims against live sources
- connect a research agent to apps like Slack or Notion via MCP

## When to choose
- you want a self-hostable, cited AI search experience
- you need agentic multi-step research rather than a single chatbot answer
- you want specialized search modes like academic, stocks, or social media
- you want to bring your own models and tools to an AI search stack

## When to avoid
- you need a simple embeddable search library rather than a full application
- you require a permissive license (it is AGPL-3.0)
- you want fully offline search without external search/model APIs
- you need enterprise-grade compliance or SLAs out of the box

## Facets
- artifact type: application
- maturity: active
- function: search-engine, agent-framework, rag, chatbot, web-scraping, mcp, llm-inference
- domain: artificial-intelligence, web-development
- platform: self-hosted
- tags: ai-search-engine, perplexity-alternative, agentic-research, citations, vercel-ai-sdk, nextjs, deep-research, agpl, search, ai-agents, retrieval-augmented-generation, natural-language-processing, web-server, nodejs, typescript

## Member repositories
- zaidmukaddam/scira (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:50.275564+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:15:16.110157+00:00, confidence not recorded.
  - readme: https://github.com/zaidmukaddam/scira (fetched 2026-08-28T04:10:50.275564+00:00, sha 81bc16f04c4b)
  - homepage: https://scira.ai (fetched 2026-08-29T08:13:28.574290+00:00, sha 3349f11ae390)
  - site_page: https://scira.ai/about (fetched 2026-08-29T08:13:28.583536+00:00, sha 85ec46756f14)
- Data as of 2026-08-30T08:39:29.467469+00:00.