Veda Developer Resources

Veda Quick Index & Predictable Endpoints

• - Public report search API: GET https://veda.ng/api/reports/search?q=agents&corpus=ai (OpenAlex-backed; see https://veda.ng/openapi.json)
• - OpenAPI 3.1 specification: https://veda.ng/openapi.json
• - MCP server over Streamable HTTP (JSON-RPC 2.0): https://veda.ng/.well-known/mcp; tools: search_essays, get_essay, search_glossary, get_glossary_term, search_reports
• - RSS feed: https://veda.ng/feed.xml
• - Sitemap: https://veda.ng/sitemap.xml
• - Agent manifests: https://veda.ng/.well-known/agents.json, https://veda.ng/ai.json, https://veda.ng/ai.txt
• - Markdown access: send Accept: text/markdown on any page URL
• - No authentication required for any read endpoint. Contact: vatsvedang@gmail.com

Veda Report Search API & REST Endpoints

Search 233,000+ indexed academic papers across the AI and Web3 corpora, backed by OpenAlex and sorted by citations. Results are cached for one hour.

GET /api/reports/search?q=stablecoin+regulation&corpus=ai&page=1&per_page=20

Parameters: q (required, min 2 chars), corpus (ai or web3), page, per_page (max 200), cursor (for cursor pagination). The full schema lives in the Veda OpenAPI 3.1 Specification.

Veda MCP Server (Model Context Protocol)

A first-party MCP server runs at /.well-known/mcp using the Streamable HTTP transport with JSON-RPC 2.0. It is stateless, needs no auth, and supports protocol versions 2025-06-18, 2025-03-26, and 2024-11-05.

Exposed tools:

• search_essays - find essays by keyword
• get_essay - full Markdown text of one essay
• search_glossary - search 100+ term glossary
• get_glossary_term - full definition of one term
• search_reports - cited academic literature via OpenAlex

curl -X POST https://veda.ng/.well-known/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize",
       "params":{"protocolVersion":"2025-06-18","capabilities":{},
                 "clientInfo":{"name":"my-client","version":"1.0"}}}'

Veda OpenAPI 3.1 Specification

The machine-readable OpenAPI specification is published at /openapi.json (with predictable aliases at /swagger.json and /docs.json). It documents 100% typed response schemas, RFC rate-limit headers, cursor pagination, batch execution, and async polling.

Veda Multi-Language SDK Packages & Tooling

Programmatic SDK packages and CLI tools for Python, TypeScript/JavaScript, and Go:

pip install vedang-cli

npm install vedang

npx @openapitools/openapi-generator-cli generate -i https://veda.ng/openapi.json -g typescript-fetch -o ./sdk

Open-source repository: github.com/vedangvatsa/vedang-website

Veda Authentication & Security Guide

All public reading and search APIs on veda.ng are keyless and require no API token. The complete WorkOS-format authentication specification is documented at /auth.md.

Veda Webhooks & Event Streaming

Live updates and event subscriptions:

• RSS Syndication: Instant updates on new publications via /feed.xml.
• SSE Live Streaming: Real-time streaming search responses supported on POST /ask with prefer.streaming: true.
• Agent Updates: Changes to tool capabilities are broadcast in /.well-known/agents.json.

Veda Markdown Content Negotiation

Every essay, every glossary term, and key pages serve Markdown to machines. Send an Accept header instead of parsing HTML:

curl -H "Accept: text/markdown" https://veda.ng/agentstack

Veda Discovery Feeds & Machine Manifests

• feed.xml - RSS feed of new essays
• sitemap.xml - every public URL
• llms.txt - structured index for LLMs (full-text version)
• agents.json - when-to-use guidance and capabilities
• ai-catalog.json - AI catalog manifest
• ai.txt - training, indexing, and citation permissions (allowed)