sales-skills/sales-frederick

Frederick AI Platform Help

Frederick AI (frederick.ai) is an AI "teammate" / co-founder for startups — a workspace of autonomous agents that execute work rather than only advising. Its differentiator vs its idea-validation siblings (aicofounder, SoloLaunch, Foundra) is that it doesn't stop at a report or a plan: Coding Agents build and ship real software (landing pages, full-stack apps, APIs, internal tools), Browser Agents drive websites (navigate, click, fill forms, extract data), Background Agents run recurring scheduled tasks (SEO content, investor updates, data collection), and Market Insights delivers AI market research, competitor tracking, customer signals, and social monitoring on a schedule. A shared Editor/workspace persists plans, research, and files as context across sessions. Work is metered in "AI Credits."

The catch for founders: Market Insights is AI-scanned research, not validated demand. Frederick can even build the smoke-test landing page (that part is genuinely useful) — but a generated market report is not proof a stranger will pay. It has no documented public API (ironic for a tool that builds APIs for you).

Step 1 — Gather context

If references/learnings.md exists, read it first for accumulated platform knowledge.

Ask only what you can't infer:

1. What do you want from Frederick?
   • A) Run Market Insights — market research, competitor tracking, customer signals, social monitoring
   • B) Use the agents — Coding (build/ship an app or landing page), Browser (automate a web task), Background (schedule recurring work)
   • C) Understand AI Credits and the free-tier gates (what the 15-credit / 2-app free plan blocks)
   • D) Get more useful output (how to frame the idea/task and manage the shared workspace)
2. What stage are you at? A raw idea, or an existing project you want to research/build/automate?
3. Is the real question "should I build this?" If so, that's a /sales-idea-validation question — route in Step 2.

Step 2 — Route or answer directly

| If the user's question is about… | Route to | |---|---| | The validate-before-building method, or comparing idea/startup-builder tools across the market | /sales-idea-validation {question} | | Building the actual landing page / smoke test as a strategy (design, copy, conversion) | /sales-funnel {question} | | Growing a pre-launch waitlist / capturing real demand signal | /sales-audience-growth {question} | | Social listening / competitor monitoring strategy across dedicated tools | /sales-social-listening {question} | | SEO / content-marketing strategy (what the Background Agent's content work feeds into) | /sales-content {question} |

When routing, give the exact command: "This is a {domain} question — run: /sales-idea-validation {original question}"

Otherwise, answer Frederick-specific questions using Step 3.

Step 3 — Frederick platform reference

Read references/platform-guide.md for the full reference — the four agent types and what each can/can't do, the Market Insights research modules, the AI-Credit metering and free-vs-Plus-vs-Pro tier gates (best-effort pricing), how to frame a task and use the shared workspace for better output, and the no-public-API automation reality.

Answer using only the relevant section — don't dump the full reference.

Step 4 — Actionable guidance

• Separate what Frederick builds from what it proves. Its agents genuinely ship things — a real landing page, a working app, a scheduled research digest. But Market Insights is AI-scanned research, not validated demand: competitor tracking and "customer signals" describe what's discussed online, not that a stranger will pay. If the real question is "should I build this?", keep Frederick's research to sharpen the pitch and route the go/no-go to /sales-idea-validation (a smoke test / pre-sale earns the decision).
• Use the Coding Agent to build the smoke test, then measure real behavior. Frederick's standout move for validation is that it can build the fake-door / smoke-test landing page itself — that's the one output that produces real signal. Have it ship the page, then drive a little traffic and read the conversion (/sales-funnel for the page strategy, /sales-audience-growth to capture signups). The built page is the proof; the market report isn't.
• Model the pricing as credit-metered tiers and flag it as best-effort. Every action spends AI Credits; name the tiers when cost comes up: a free tier (~15 credits, 1GB, 2 apps, limited Market Insights, no recurring background agents) → Plus (~$25/mo, ~1,000 credits, 10GB, unlimited apps, full Market Insights, recurring background agents) → Pro (~$50/mo, ~2,500 credits, 30GB), annual ~24% off. Sources conflict (some third-party listings quote ~$8/$16 — likely stale/promo), so tell the user to confirm live at frederick.ai/pricing. Frame value as "how many credit-heavy agent runs per month" vs the monthly fee.
• There's no public API — don't design an integration around it. Despite building APIs for you, Frederick exposes no documented public API, webhooks, MCP, or Zapier/Make. If asked to automate/export, say so plainly: copy outputs from the workspace manually, or (for a validation pipeline) automate the real signal — landing-page analytics / waitlist events — not the generated research. Use the Background Agent for recurring work inside Frederick; it isn't an outbound integration surface.
• Frame the task richly and lean on the shared workspace. Agent output tracks input depth — give a specific customer, the exact problem, and the differentiation, and store plans/research/files in the Editor so agents carry context between runs. Treat every generated asset (page, plan, report) as an editable first draft, and re-run variants.

If you discover a gotcha or tip not in references/learnings.md, append it there with today's date.

Gotchas

Best-effort from research (2026-07) — Frederick's agent lineup, credit costs, and pricing move fast; verify at frederick.ai.

• Market Insights is research, not demand. Competitor tracking and "customer signals" report what's discussed online; they don't prove anyone will pay. Pair with a real demand test (/sales-idea-validation).
• Credits meter everything. Agent runs (coding, browser, background, insights) all draw down AI Credits; the free tier's 15 credits and 2-app cap run out fast, and recurring background agents are gated to paid tiers. Watch the credit balance before scheduling recurring work.
• No documented public API / webhooks / Zapier. It's a UI workspace — don't plan an external integration around it. The Background Agent automates work within Frederick, not out to your stack.
• Pricing sources conflict. Homepage vs third-party listings disagree (e.g. ~$25/$50 vs ~$8/$16) — treat all figures as best-effort and confirm live at frederick.ai/pricing.
• A shipped app isn't a validated app. Frederick will happily build a full product for an idea nobody wants. Building is cheap now; the scarce thing is evidence of demand — get that first.
• Output tracks input depth. A thin, one-line task yields generic output; frame a specific customer + problem + differentiation and use the shared workspace for context.

Related skills

• /sales-idea-validation — The tool-agnostic validate-before-building method + the full validator/builder landscape (use this to actually decide build-or-not; Frederick is one option among aicofounder, SoloLaunch, IdeaProof, Preuve, …)
• /sales-aicofounder — aicofounder — a close sibling AI co-founder with phased idea→launch workflow + grounded Reddit/X demand research
• /sales-sololaunch — SoloLaunch — a sibling AI startup builder that generates a "startup foundation" (names, logo, timeline, marketing strategy)
• /sales-funnel — Build the smoke-test / landing page the demand test runs on (Frederick's Coding Agent can ship it; this is the page strategy)
• /sales-audience-growth — Grow a pre-launch waitlist and capture the real demand signal
• /sales-social-listening — Competitor / social monitoring strategy across dedicated tools (what Market Insights overlaps)
• /sales-content — SEO / content-marketing strategy (what Frederick's Background Agent content work feeds)
• /sales-do — Not sure which skill to use? The router matches any sales objective to the right skill. Install: npx skills add sales-skills/sales --skill sales-do -a claude-code

Examples

Example 1: "Frederick's Market Insights says the market's great — am I validated?"

User says: "Frederick ran competitor tracking and market analysis on my idea and it looks promising. Good to build?" Skill does: Separates Frederick's AI-scanned research from validated demand — competitor tracking and "customer signals" describe what's discussed online, not that a stranger will pay. Keeps the research to sharpen the pitch (competitors, positioning, risks) and routes the go/no-go to /sales-idea-validation: use Frederick's Coding Agent to build the smoke-test landing page, then measure real conversions. Result: The founder gets a decision from stranger behavior, using Frederick for the research and the page, not the verdict.

Example 2: Have Frederick build and schedule the launch work

User says: "Can Frederick build my landing page and then keep researching my competitors every week?" Skill does: Confirms the Coding Agent can ship the landing page and the Background Agent can run recurring weekly Market Insights (competitor tracking) on a schedule — but flags that recurring background agents are gated to paid tiers and that every run spends AI Credits, so watch the balance. Points the page's conversion strategy to /sales-funnel and frames the weekly research as monitoring, not demand proof. Result: A concrete plan — build once, schedule the research — with clear credit/tier expectations.

Example 3: Can I automate Frederick via API? (developer/automation)

User says: "I want to trigger Frederick's agents and pull the research out from a script — what's the API?" Skill does: States plainly there's no documented public API, webhooks, MCP, or Zapier/Make (ironic, since Frederick builds APIs for you), so there's no supported way to trigger agents or export research programmatically — copy outputs from the workspace manually, or automate the real signal (landing-page analytics / waitlist events) instead. Notes the Background Agent handles recurring work inside Frederick, not outbound integration. Result: The user avoids building on a non-existent API and picks a workable path.

Troubleshooting

Ran out of credits / can't schedule recurring research

Symptom: Agents stopped, or a recurring Background Agent won't run. Cause: Every agent run spends AI Credits; the free tier caps at ~15 credits and 2 apps, and recurring background agents are gated to paid tiers. Solution: Check the credit balance and tier — full Market Insights and recurring background agents need Plus/Pro. Weigh how many credit-heavy runs you'll do per month against the monthly fee; confirm current credits/limits at frederick.ai/pricing (treat figures as best-effort).

Output feels generic / like plain ChatGPT

Symptom: The generated research, plan, or page could fit any product. Cause: The task input was thin, and the agents lacked shared context. Solution: Re-run with a rich brief — specific customer, exact problem, differentiation, current stage — and store plans/research/files in the Editor workspace so agents carry context between runs. Compare variants.

"Is there an API to automate this?"

Symptom: Looking for API docs, webhooks, or a Zapier connector. Cause: Frederick is a UI workspace with no documented public API/webhooks/Zapier; the Background Agent automates work within Frederick, not out to your stack. Solution: Don't build an integration around it — copy outputs manually. For a validation pipeline, automate the real demand signal (landing-page analytics, waitlist events) instead. See /sales-idea-validation.