Connect AI Agents, IDEs & Custom GPTs
Connect ChatGPT, OpenAI Codex, Claude Desktop, Cursor, Antigravity, Windsurf, Cline, or custom agent runtimes directly to FullStack AI Hub. Equip your AI assistants with 560+ verified concepts, 27 tools partitioned across Knowledge Consumer and Curator personas, 15 REST action endpoints, 9 resources, 5 interactive UI widgets, official engineering skills, live AST codebase graphs across reference architectures (vLLM, Chatwoot, LangGraph, DeepAgents), 13 WebAssembly runtime challenges, 11 seminal paper teardowns, and Reciprocal Rank Fusion semantic search.
Official manifest for ChatGPT Plugin discovery, OAuth metadata, and API specifications.
Role-filtered OpenAPI 3.1 specification for ChatGPT Custom GPTs with 10 read-only GET operations.
Authoring OpenAPI 3.1 specification with Bearer security scheme and 5 POST/GET operations.
https://fullstackaihub.com/mcphttps://fullstackaihub.com/ssehttps://fullstackaihub.com/mcp/healthhttps://fullstackaihub.com/gatewayhttps://fullstackaihub.com/audit-loghttps://fullstackaihub.com/gateway/routeshttps://mcp.fullstackaihub.com/mcphttps://mcp.fullstackaihub.com/ssehttps://mcp.fullstackaihub.com/healthAgent Configuration Generator
Dynamic replacement: automatically updates the Authorization header snippet below.
Instant access via apex domain with Traefik priority 100 bypass. Zero DNS propagation delay.
{
"mcpServers": {
"fullstack-ai-hub": {
"url": "https://fullstackaihub.com/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}~/Library/Application Support/Claude/claude_desktop_config.json%APPDATA%\Claude\claude_desktop_config.json~/.config/Claude/claude_desktop_config.jsonNative desktop assistant by Anthropic. Connects over Streamable HTTP. Restart your agent or client after saving this configuration.
Live Interactive Dual-Protocol Sandbox
Retrieve dynamic 3-9 layer AST codebase graph, nodes, classes, functions, and dependency edges.
Injected into cURL & HTTP request.
curl -X POST "https://fullstackaihub.com/mcp" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "fsai.get_codebase_graph",
"arguments": {
"systemId": "inside-vllm",
"layerId": "memory-kvcache",
"includeSnippets": true
}
}
}'import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport }
from "@modelcontextprotocol/sdk/client/streamableHttp.js";
const client = new Client({ name: "my-agent", version: "1.0" });
// Instant Zero-DNS Gateway (Recommended):
const transport = new StreamableHTTPClientTransport(
new URL("https://fullstackaihub.com/mcp"),
{ requestInit: { headers: {
Authorization: "Bearer YOUR_API_KEY"
}}}
);
// Or via Dedicated Subdomain:
// new URL("https://mcp.fullstackaihub.com/mcp")
await client.connect(transport);
// Query dynamic AST codebase graph
const astGraph = await client.callTool({
name: "fsai.get_codebase_graph",
arguments: { systemId: "inside-vllm", includeSnippets: true }
});In production mode, all MCP requests require a Bearer token in the Authorization header.
/health endpoint is publicly available without credentials for Traefik and Docker health checks.Rate limit: 30 requests per minute per client IP address. Breaching the threshold yields an HTTP 429 Too Many Requests response with a Retry-After header.
RESTful Custom GPT Actions (15)
Callable directly via standard HTTP requests or OpenAI Custom GPT Actions with OpenAPI 3.1 schema.
/api/consumer/searchSemantic & full-text search across concepts, systems, and papers
/api/consumer/concepts/:slugCanonical concept lesson by slug with mental models and sources
/api/consumer/concepts/:slug/relationsDirected knowledge graph relations (prerequisites, enables)
/api/consumer/systems/:slugArchitectural system breakdown with dynamic layers and hot-path code
/api/consumer/systems/:slug/traceGranular execution trace steps, packet flows, and state diffs
/api/consumer/codebase/:systemIdDynamic 3-9 layer AST codebase graph and symbol nodes
/api/consumer/papersList academic research paper breakdowns with verified citations
/api/consumer/papers/:slugRetrieve full research paper breakdown by slug with mathematical formulas
/api/consumer/wasm-challengesList interactive WebAssembly AI runtime kernel challenges
/api/consumer/wasm-challenges/:idInteractive Pyodide WebAssembly challenge definition and assertions
/api/curator/validateContent schema validation, Acorn MDX LaTeX math checks, and referential integrity audit
/api/curator/conceptCurate or update a canonical concept lesson MDX in the knowledge graph
/api/curator/relationPropose a directed knowledge graph relation with cycle prevention
/api/curator/assessmentGenerate interactive assessment YAML with theory and decision scenarios
/api/curator/healthAuthorization health check returning curator service status and granted scopes
Available Tools Catalog (27)
Callable via standard JSON-RPC 2.0 tools/call method, segmented by persona and privilege tier.
fsai.get_codebase_graphRetrieve dynamic 3-9 layer AST codebase graph, nodes, classes, functions, and dependency edges
fsai.query_codebase_symbolsSearch classes, functions, modules, and language concepts across reference architectures
fsai.search_knowledgeFull-text search across concepts, systems, guides, and paths
fsai.get_hybrid_contextDense Vector + BM25 semantic RAG retriever with RRF fusion
fsai.get_conceptFull canonical concept lesson by slug with mental models, failure modes, and sources
fsai.get_concept_relationsDirected knowledge graph edges for a concept
fsai.get_system_breakdownComplete 7-layer architectural system breakdown with paper mappings and hot-path code
fsai.get_system_step_traceGranular execution traces, packet flows, and state variable deltas
fsai.query_architecture_routeShortest directed route between two system components
fsai.get_learning_pathStructured prerequisite-ordered learning path
fsai.get_evidence_sourcePrimary evidence source metadata and citations
fsai.get_assessmentInteractive assessment questions and scenarios for a concept
fsai.get_lab_scenarioDeterministic lab scenario configurations and rubrics
fsai.list_territoriesComplete curriculum taxonomy hierarchy
fsai.get_publicationResearch paper breakdowns, incident postmortems, and technical essays
fsai.list_papersList canonical academic paper breakdowns from library with research methods and verified sources
fsai.get_paper_breakdownRetrieve full research paper breakdown by slug with mathematical formulas and limitations
fsai.list_wasm_challengesList interactive WebAssembly AI runtime kernel challenges covering autograd, attention, and memory
fsai.get_wasm_challengeRetrieve complete WebAssembly AI runtime kernel challenge by ID with starter code and assertions
fsai.get_reference_guideProduction AI reference architecture synthesis guide
fsai.curate_conceptAuthor or update a canonical concept lesson MDX with frontmatter and mental models
fsai.propose_concept_relationPropose or validate a directed knowledge graph relation between concepts with cycle prevention
fsai.generate_assessmentGenerate interactive assessment YAML with theory questions and decision scenarios
fsai.validate_contentSchema validation, Acorn MDX LaTeX math syntax checks, and referential integrity audit
fsai.inspect_curriculum_healthQuantitative health metrics for the knowledge graph, domain distributions, and orphan counts
fsai.get_server_profileOpenAI plugin profile, server metadata, dynamic layer support, and skill catalog
fsai.get_audit_logQuery the append-only cryptographic tool invocation audit log and zero-trust policy decisions
Interactive MCP Apps UI Widgets (5)
Lightweight visual exploration widgets embedded directly inside ChatGPT, Claude, and MCP host applications via _meta.ui.resourceUri and openai/outputTemplate.
fsai.get_wasm_challengeui://wasm-kernel-runnerfsai.get_system_step_traceui://system-tracerfsai.get_conceptui://concept-graphfsai.get_codebase_graphui://codebase-graphfsai.get_assessmentui://assessment-quizMCP Resources (9)
Readable via resources/read with direct URI addressing.
skill://fsai/{skill}/SKILL.mdOpenAI ChatGPT official skill definitions with SHA-256 integrity digestsfsai://atlasComplete curriculum atlas with all territories, domains, and conceptsfsai://concepts/{slug}Individual concept lesson by slug with learning outcomes and failure modesfsai://systems/{id}7-layer system breakdown and architecture data by IDfsai://codebase/{systemId}Dynamic multi-layer AST codebase graph and symbol nodes by system IDfsai://evidence/{sourceId}Primary evidence source citations and metadata by IDfsai://paths/{slug}Learning path curriculum and milestones by slugfsai://papers/{slug}Seminal academic research paper breakdowns with mathematical formulation and source citationsfsai://wasm-challenges/{id}Interactive WebAssembly AI runtime kernel challenge definition, starter code, and assertionsMCP Prompts (3)
Interactive prompt templates callable via prompts/get.
teach-conceptStructure a production engineering lesson for a conceptaudit-architectureCompare a proposed architecture against reference breakdownscompare-systemsSide-by-side architectural comparison of two AI systemsfsai.get_wasm_challenge.fsai.get_paper_breakdown.