Installation — Quick Start
Create your account, issue an API key, and connect your first MCP client in about 15 minutes.
This guide walks you from zero to a working MCP connection. Estimated time: 15 minutes.
1. Create an account
- Open Get started free.
- Sign in with email magic link or Google.
- You land on Dashboard with a default team.
By continuing, you agree to our Terms of Service and Privacy Policy.
2. Create an API key
- Go to Dashboard → API Keys.
- Click Create key and copy the secret (
sk_live_…). - Store it in a password manager or host env var (e.g.
SMARTGATE_API_KEY). Do not commit it to git.
Free plans include up to 2 API keys per team.
3. Open Connect
- Go to Dashboard → Connect.
- Choose your MCP host (Cursor, Claude Desktop, Windsurf, Hermes, OpenClaw, or Generic).
- Copy the JSON snippet — it includes the correct URL (
POST /api/mcp, Streamable HTTP) and Authorization header for your deployment.
Production typically uses:
https://smartgate.network/api/mcpLocal development uses the Next.js BFF at http://localhost:3000/api/mcp, or direct Python at http://localhost:8000/mcp/; Connect shows the right value for your environment.
4. Paste into your MCP client
- Cursor: merge into
~/.cursor/mcp.jsonundermcpServers, then restart Cursor. - Claude Desktop: merge into
claude_desktop_config.json, then restart the app. - Other hosts: follow the host-specific field names on the Connect page.
See the Connect Guide for per-host details.
5. Verify the connection
In your MCP client, confirm the server shows as connected, then invoke:
tools/list— you should see all 7 SmartGate toolstools/callwithsmart_fetchon a test URL
If the client reports auth errors, double-check the Bearer token and URL from Connect.
6. Optional — REST audit headers
For multi-step agent tasks over REST (not MCP), add optional headers documented in MCP Endpoint → Audit headers. MCP calls tag route=mcp automatically.
Next steps
- Connect Guide — host-specific setup
- MCP Endpoint — protocol details
- Security — key rotation and retention
- Dashboard Playground — test fetch/compress via REST (separate from MCP)
What installed means for an MCP gateway
SmartGate is a hosted gateway. Quick Start is not a local binary install: you create an account, mint an API key, and point an MCP client at Streamable HTTP. Success looks like initialize completing, ools/list returning the seven smart_* tools, and at least one audited tool call in Activity Logs.
Recommended first hour
- Sign up and verify email if prompted.
- Create a personal or team workspace; note the plan entitlements on Pricing.
- Mint an API key; store it in your secret manager—never commit it to git.
- Open Dashboard Connect; copy host JSON (Connect).
- Restart the host; confirm tools populate.
- Run a bounded Playground preset or a single MCP tool call—not an unbounded agent loop on Free RPM.
- Open Activity Logs and confirm eam_id, tool name, and oken_used.
Fair-use and upgrades
Free and Playground use strict fair-use rate limits. Bursts may soft-fail with HTTP 429, a limit_scope, and Retry-After when available. That is expected entitlement behavior for individual evaluation, not a multi-user load profile. Upgrade when your team needs higher MCP RPM, shared workspace limits, longer audit retention, or optional L2 hard cap behavior described in billing docs.
Failure modes to expect early
| Symptom | Likely cause | Next step |
|---|---|---|
| 401 on /api/mcp | Missing/wrong Bearer token | Regenerate key; update Connect JSON |
| Empty tools | Handshake incomplete | Restart host; confirm Streamable HTTP |
| 429 under light use | Fair-use RPM or daily cap | Slow retries; upgrade if shared load |
| No audit row | Different team/key than dashboard filter | Align team scope in Logs filters |
When installation is done, read Tools for when to search vs fetch vs pipe, and Security before inviting teammates.
Practical adoption notes
Treat SmartGate as shared infrastructure: one gateway URL, team-scoped keys, and audit on by default. Prefer pipe templates for repeatable research paths so Activity Logs show a coherent chain instead of ad-hoc tool spam. Keep host-model prompts focused on decisions; let Research and Context primitives prepare evidence. When FinOps asks for savings, export Usage Reports alongside the gateway benchmark explanation—never present Est. USD saved as a vendor LLM invoice. When security asks for evidence, export an audit window and show retention tied to the current plan catalog.
If you are migrating from scripts, replace raw HTTP fetch helpers first, then add dedup and compression, then budget guard. If you are migrating from a model gateway mindset, stop expecting chat completions on /api/mcp and wire MCP hosts instead. If you are writing human-facing product copy, lead with Smart Research / Smart Decision outcomes; keep MCP JSON on builder paths. These habits keep the site and the product aligned with the Positioning Bible: intelligence layer between AI and the world, not an LLM proxy.
Related reading
- Overview for the capability map and plan table
- Connect Guide for host JSON and troubleshooting
- Tools Reference for when to search, fetch, compress, or pipe
- Activity Logs for audit, trace, and export
- Pricing for fair-use limits, Gainshare, and plan ceilings
- Token Control, Traffic Customs, and Audit and Compliance for pillar narratives aligned with this docs set
Keep English-first public docs consistent with the Positioning Bible: SmartGate is an MCP-native algorithmic gateway and intelligence layer—not a chatbot, ChatGPT alternative, AI search engine, LLM proxy, or model router. Humans research and decide through Web experiences; agents connect through MCP/API; both share the same runtime primitives and team-scoped governance.
One more operational reminder
Document your team naming scheme, key rotation cadence, and who owns Gainshare questions before you scale agent count. Prefer bounded Playground or single-tool MCP checks when validating Connect JSON. Prefer audited smart_pipe templates for recurring research once the handshake works. Prefer exporting Activity Logs before plan downgrades that shorten retention. These habits cost little and prevent the two failure modes we see most: empty tool lists from incomplete MCP initialize, and missing audit evidence when a reviewer asks what an agent fetched last week.