Plug in, in a minute.
Docs
JustOne Connector is a standard remote MCP server (Streamable HTTP + OAuth 2.1) — any MCP-capable client can connect.
Connector URL
https://justoneconnector.com/api/mcpConnecting each client
Claude (web / desktop / mobile)
- Open Settings → Connectors → Add custom connector
- Paste the connector URL above and click Add
- Complete the OAuth login (email one-time code) and authorize
- In a new chat open + → Connectors → Tool access → choose Loaded tools; Claude has no direct link to this setting, and future chats will then default to loading One
- Now ask Claude to query data, search or call models right in the chat
ChatGPT (Pro / Business / Enterprise)
- Settings → Apps → Advanced → enable Developer mode
- Add a connector, paste the connector URL, choose OAuth
- Sign in, authorize, and use it in any conversation
Note: free ChatGPT doesn't support custom connectors yet (an OpenAI platform limitation).
Claude Code
claude mcp add --transport http justone https://justoneconnector.com/api/mcpThe first tool call opens a browser for the OAuth login. An API key works too:
claude mcp add --transport http justone https://justoneconnector.com/api/mcp --header "Authorization: Bearer jc_sk_YOUR_KEY"Codex CLI
codex mcp add justone --url https://justoneconnector.com/api/mcpCursor / Windsurf / other MCP clients
Add this to mcp.json:
{
"mcpServers": {
"justone": {
"url": "https://justoneconnector.com/api/mcp"
}
}
}Agent frameworks / scripts (API key)
Create an API key in the dashboard and send it in the request header — no OAuth needed:
POST https://justoneconnector.com/api/mcp
Authorization: Bearer jc_sk_xxxxxxxx
Content-Type: application/json
Accept: application/json, text/event-stream
{"jsonrpc":"2.0","id":1,"method":"tools/list"}The OpenAI Agents SDK, Vercel AI SDK and LangChain all mount remote MCP servers with this config.
Tool reference
| Tool | Description | credits/call |
|---|---|---|
| search_social_ecommerce_kol_data | Front door for Pgy, Xingtu, WeChat and 30+ native social/commerce data sources; finds the right endpoint before execution | free |
| find_social_user | Find users by name in native Xiaohongshu, Douyin, Kuaishou or TikTok account indexes — not Google | 3–5 |
| get_user_posts | Read posts, notes, videos or articles published by one known account across 12 platforms, with pagination | 3–6 |
| social_search | Agentic social search: hand it a natural-language intent; the server plans short query variants, fans out in parallel, spam-scores candidates on X's open-sourced ranking signals, and a model curates authentic voices — returned with a search log and itemized cost | metered + 5 fee, budget-capped (default 40) |
| social_search_raw_unfiltered | Raw single-platform search page: unfiltered, engagement-ranked, top results often promotional; for pagination and exact listing lookups only. Twitter rows carry a _quality bait score | 3–23 |
| get_item_detail | Product details (Amazon / Shopee / Taobao / JD / Xianyu / 1688…) | 3–15 |
| get_content_detail | Post / video / article details | 3–6 |
| get_comments | Comments and product reviews, with pagination | 3–6 |
| get_user_profile | User / creator profile info | 3–5 |
| hot_list | Trending charts (TikTok / RED / Douyin / Weibo / Douban) | 3–6 |
| resolve_share_url | Turn share links into canonical IDs | 3–5 |
| list_data_apis / call_data_api | Browse and call all 280+ raw endpoints (Twitter, LinkedIn, Reddit, IMDb, Pgy, Xingtu…), each individually priced | per catalog |
| list_models / ai_chat | Chat with 160+ models, per-model rates, billed by tokens | per model |
| model_committee | Several models answer one question in parallel (Pro and up) | sum of models |
| humanize | Rewrite AI-generated text to read human, in-house style guide on claude-opus-4.6 | by tokens |
| deep_think | One hard question, deeply reasoned advice from gpt-5.5-pro or claude-fable-5 (effort maxed) | by tokens |
| generate_image(_batch) | Text-to-image / reference-guided edits, multi-variant, get_cost previews (Starter and up) | ≈12/image |
| generate_video(_batch) | Text-to-video: Seedance 2.0 / 1.5 Pro / Veo 3.1, async jobs + jobs_wait (Pro and up) | 8–80/sec |
| get_generation_result / jobs_wait / show_generations | Check one async job, wait on a batch, or review recent generations (rendered as gallery cards in MCP Apps clients) | free |
| analyze_media | Vision-model analysis of images / videos / PDFs / audio (the base for virality and scene breakdowns) | by tokens |
| transcribe_audio | Speech-to-text, billed on actual duration (Starter and up) | 1/min |
| generate_speech | Text-to-speech, returns a hosted audio URL (Starter and up) | ≈8/1k chars |
| research_search | Academic & knowledge sources: arXiv / OpenAlex / Semantic Scholar / Crossref / Wikipedia / Open Library / Gutenberg, normalized output | 1 |
| open_data | Authoritative statistics: 81 World Bank indicators + Our World in Data | 1 |
| hn_search | Hacker News full-text search and threads | 1 |
| gov_data | Government open-data portals (Macau data.gov.mo / Hong Kong data.gov.hk / US catalog.data.gov): search datasets, list resources, fetch files or invoke realtime APIs (search/detail 2, fetch 4). Plus the exclusive OpenMO source: full-text Macau government records — Official Gazette, contracts, tenders, Legislative Assembly, court judgments, land grants, personnel (search/detail 4, full text 8) | 2–8 |
| media_import_url / list_media | Import remote media into your library as generation/analysis references; list_media reviews the library for free | 1+ |
| sandbox_exec / sandbox_file / sandbox_extend / sandbox_kill | Isolated cloud sandbox for Python/Bash/JS: pip/npm, network, state persists across calls, extendable lifetime, charts come back as hosted URLs (paid plans; 40% launch discount) | 0.6/min |
| deploy_site / site_info / delete_site | Publish a static site to your-slug.justoneconnector.app (Cloudflare's global edge, up to 100 files / 20MB, SPA routing supported); redeploying the same slug updates it, hosting is free within your plan's site limit (Pro and up) | 10/deploy |
| brand_research_start / brand_generate_doc / get_brand_docs | Brand research: crawl a site into a corpus → six marketing strategy documents (product info / competitors / brand voice / marketing strategy / content strategy / llms.txt), reusable across sessions | pages + tokens |
| get_workflow_instructions / get_workflow_bundle_file | Workflow catalog: 8 multi-step playbooks — product research, competitor analysis, brand strategy, KOL scouting, social listening, SEO audit, content multiplication, virality breakdowns — agents follow the loaded SKILL | free |
| search_web | Google keyword search across 22 verticals: web / news / images / shopping / maps / scholar / researcher profiles / patents / trends / jobs / hotels… (Starter and up) | 1–2 |
| google_lookup | Detail lookups by ID: place reviews / photos / posts, product details, scholar author pages, citation formats, patent details, Google Lens | 1–2 |
| fetch_webpage | Any webpage as Markdown / HTML / JS-rendered | 1–4 |
| account_status | Plan, balance and recent usage | free |
Every tool result is annotated with its actual cost; use list_data_apis to see the exact price of any raw endpoint. In MCP Apps-capable clients (like Claude), the generation gallery, account overview and brand documents render as interactive widgets.