# superagnt_ — Agent workflows through one MCP server and REST surface.

> Source: AgentsIndex — https://agentsindex.ai/superagnt (structured, researched, re-verified)
> Facts last verified: 2026-10-08

| Fact | Value |
| --- | --- |
| Website | https://superagnt.com/agent-toolkit |
| Price | from $49/month |

## What is superagnt_?

Superagnt is an agent workflow platform for operators, growth teams, founders, developers, and RevOps teams that routes work through one MCP server and the same REST surface. It combines data_* and agnt_db_* tools, a database, canvas, and knowledge + memory so agents can fetch data, write rows, and continue workflows. It integrates with Claude Desktop and supports custom MCP servers. Plans run Starter $49/mo, Growth $149/mo, Scale $399/mo, and Enterprise custom.

## What's bundled behind one key

The toolkit's own docs reference lists 10 data sources across 257 endpoints reachable over MCP or plain REST with one API key and one credit balance (superagnt.com/docs). The homepage enumerates nine of them by name — Instagram, TikTok, Facebook, YouTube, LinkedIn, Reddit, X, generic web scraping/search, and a people+company enrichment source (email finding, verification, mobile lookup, firmographic data) — each returning structured JSON instead of raw HTML. Past the data layer, the same connection gives an agent a real workspace Postgres database (SQL, migrations, CSV loads, 10 GB to start per workspace), shareable live-query dashboards ("canvas"), Slack/iMessage bindings, Gmail/Outlook inbox triage, inbound and outbound signed webhooks, scheduled and queued jobs, bulk data pipelines for processing large row sets outside the chat context, and the ability to mint scoped, per-agent MCP servers so a researcher agent and an outreach agent don't share the same tool surface. This is a wider bet than a typical scraping API: it's trying to be the backend an agent needs instead of just a data source (superagnt.com).

## How the pricing actually works

Three self-serve plans, verified on the pricing page: Starter $49/mo ($20 of data credits included, 1 workspace, 3 deployed agents, 10 schedules, 3 queues, a 10-table/25,000-row database, 60 requests/min), Growth $149/mo ($90 credits, 3 workspaces, 15 agents, 300 req/min), Scale $399/mo ($300 credits, 10 workspaces, 50 agents, 600 req/min), plus custom Enterprise. Every plan gets every tool and data source — the plans differ only in volume and throughput. One credit equals one cent, split into two wallets: a data wallet the plan tops up monthly for API calls, and a separate AI wallet you fund yourself in the dashboard to pay for the tokens a *deployed* agent burns (your own coding agent, e.g. Claude Code or Cursor, isn't metered this way — it brings its own model). The page states AI credits are purchased separately but doesn't publish a $/token conversion rate anywhere on this page. Billing is per successful call only — a failed request isn't charged — and every cap is hard: cross one and the agent gets an upgrade offer rather than an overage bill. A concrete example: a single Instagram profile lookup costs $0.002, i.e. 0.2 credits (superagnt.com/tools/instagram). There's a no-card trial: $3 of data credits at Growth-tier limits, with the 7-day clock starting at your 25th billable data call, first database use, or when the $3 runs out — not at signup.

## What's open, what isn't

Setup runs through a GitHub-hosted plugin marketplace — github.com/superagnt/plugins (MIT licence, 4 stars) — which installs the MCP server connector and matching Claude Code skills; a companion repo, leverage (MIT, 3 stars), holds blueprint workflows. Both are small and recently active (last commits early October 2026) but they are the install/glue layer, not the product: the scraping engines, enrichment providers, database, and automation backend run as closed-source SaaS reached only through the hosted mcp.superagnt.com endpoint. The GitHub organization lists no public members (github.com/superagnt, checked 8 October 2026).

## How young this is

superagnt.com has no snapshots in the Wayback Machine before 19 August 2026 (the earliest capture is static JS/CSS assets from that date), putting the public site at roughly seven weeks old as of this review. There's no /about page (404), no /changelog, and no /status page. The Terms of Service name only "superagnt" and set Delaware as the governing law, but no legal entity name (LLC, Inc., Corp.) appears anywhere on the public site, and the GitHub org has no listed members. None of this is unusual for an early-stage product, but it means there's no founder name, operating history, or independent review on G2, Product Hunt, or elsewhere to weigh this listing against — a buyer is assessing the product pages themselves, not a track record.

## Security and data handling, as published

Dedicated /security and /privacy pages exist (both last updated 5 July 2026) and go beyond boilerplate: the workspace Postgres is provisioned per workspace with no connection string handed out, memory and data don't cross workspace boundaries, connected-tool credentials use scoped OAuth flows, and Customer Data is stated not to be used to train third-party foundation models unless a customer explicitly enables a feature requiring it. There's a stated vulnerability-reporting process and an incident-response commitment. No compliance certification (SOC 2, ISO 27001, etc.) is claimed; the security page says documentation "may be available for customers evaluating paid or higher-risk deployments" on request, which reads as not yet achieved rather than available today.

## How much does superagnt_ cost?

| Plan | Price | What's included |
| --- | --- | --- |
| Starter | $49/mo | 1 workspace · 3 deployed agents; 10 schedules · 3 queues; Database: 10 tables, 25,000 rows; 1 inbox · 1 messaging line; 60 requests / min; Every tool, data source and connection |
| Growth | $149/mo | 3 workspaces · 15 deployed agents; 50 schedules · 15 queues; Database: 50 tables, 250,000 rows; 3 inboxes · 3 messaging lines; 300 requests / min; Every tool, data source and connection |
| Scale | $399/mo | 10 workspaces · 50 deployed agents; 250 schedules · 50 queues; Database: 250 tables, 2,500,000 rows; 10 inboxes · 10 messaging lines; 600 requests / min; Every tool, data source and connection |
| Enterprise | Custom | Past Scale. Limits and throughput sized to your volume, with SSO, invoicing and an SLA. |

## Frequently asked questions

**What is superagnt_?**

Superagnt is an agent workflow platform for operators, growth teams, founders, developers, and RevOps teams that routes work through one MCP server and the same REST surface. It combines data_* and agnt_db_* tools, a database, canvas, and knowledge + memory so agents can fetch data, write rows, and continue workflows. It integrates with Claude Desktop and supports custom MCP servers. Plans run Starter $49/mo, Growth $149/mo, Scale $399/mo, and Enterprise custom.

**How much does superagnt_ cost? Is it free?**

Superagnt_ has 4 paid plans: Starter at $49/mo, Growth at $149/mo, Scale at $399/mo.

**What is superagnt_ used for? Who is it for?**

Superagnt_ is used for MCP + REST, data layer, and database. It's built for Operators, Growth teams, and Founders.

**Does superagnt_ have an API and what does it integrate with?**

The product exposes MCP and plain REST on the same surface, with tool endpoints like data_* and agnt_db_*. It integrates with Slack, Gmail, Outlook, Stripe, GitHub, and 25 more.

**Our take:** Check the workspace and database ceilings before committing: Starter allows 1 workspace and 10 tables, Growth 3 workspaces and 50 tables, and Scale 10 workspaces and 250 tables. If your workflow needs more than 2,500,000 rows or 600 requests per minute, you are into Enterprise territory.

## Coverage elsewhere

| Source | Kind | Read |
| --- | --- | --- |
| web.archive.org | Independent source | undated |

## Corrections

If you represent superagnt_, or anything here is wrong or out of date, send a correction with a source. The operations this site accepts from an agent are listed at https://agentsindex.ai/.well-known/agents.json — that catalog needs no key to read, and each entry states its own auth.