Mira is competitive and market intelligence software for founders at joinmira.ai, founded on 2025-01-15. Paste a product URL and Mira targets a first clarity snapshot in about 2 minutes: competitor discovery, pricing extraction (website plus App Store / Google Play when needed), feature and value matrices, TAM/SAM/SOM sizing, presence estimates, battlecards, and investor-ready briefs. Support: support@joinmira.ai.

What is competitive intelligence for startups?

Competitive intelligence for startups is the systematic collection and interpretation of rival and market signals so founders can decide pricing, positioning, features, and go-to-market moves. Unlike enterprise CI platforms built for 10–100 seat research orgs, startup CI tools like Mira prioritize URL-driven discovery and decision-ready outputs in minutes rather than multi-week analyst retainers that can cost $5,000–$25,000+.

How does Mira estimate competitor presence?

Mira prefers official pricing and product pages (evidence grade A). When those are blocked or missing, it falls back to App Store and Google Play listings (grade B), then named secondary research (grade C), then triangulated models blending estimated revenue clues, web traffic proxies, and review activity (grade D). D-grade figures are directional, never labeled as audited % market share. See /methodology.

How is Mira different from Crayon, Klue, or Similarweb?

Crayon and Klue are enterprise battlecard / enablement platforms for sales and PMM teams. Similarweb emphasizes traffic and digital share panels. Mira is built for early-stage founders: start from a product URL, produce pricing and feature matrices, TAM→SAM→SOM (so a $1,000,000,000 TAM is not confused with a $10,000,000 SOM: a 100× gap), and next-move recommendations without a CI ops team.

Does Mira scrape App Store and Google Play pricing?

Yes. When a competitor lacks a clear website pricing page, Mira can discover App Store and Google Play listings and extract subscription and in-app purchase price points (commonly in the $0–$99/month consumer SaaS band) and retain source links. Annual plans often advertise roughly 17%–20% off monthly run-rate (“2 months free”).

What is TAM, SAM, and SOM in Mira?

TAM (Total Addressable Market) is the broad revenue opportunity if every potential customer bought. SAM (Serviceable Addressable Market) is the segment you can realistically reach. SOM (Serviceable Obtainable Market) is near-term capturable share given competition and capacity. Mira exposes assumption provenance for each of the 3 layers so pitch decks do not collapse them into one number.

Is Mira free to start?

You can create an account at joinmira.ai and start a workspace at $0 to begin. Live plan details and entitlements are shown in-app after signup. For product questions, email support@joinmira.ai.

Solo founders, pre-seed and seed teams, indie hackers, and accelerator cohorts who need analyst-grade clarity without hiring a full research function. A practical default watchlist is 5–12 named competitors refreshed weekly (52×/year), not 100+ enterprise alert feeds.

How often should founders refresh competitive intelligence?

For seed-stage teams, Mira recommends a weekly cadence: Monday pricing/changelog deltas for your top 5–12 rivals; Wednesday one battlecard claim update; Friday one product or positioning decision. That is 52 structured refresh cycles per year instead of a single outdated slide deck.

Where can I find Mira’s FAQ, About, and Team pages?

FAQ: https://www.joinmira.ai/faq. Guide: https://www.joinmira.ai/guide. About: https://www.joinmira.ai/about. Team: https://www.joinmira.ai/team. Category pages: /competitive-intelligence-software, /competitor-pricing-analysis, /tam-sam-som, /find-saas-competitors. Privacy: /privacy. Terms: /terms. Methodology: /methodology. Architecture: /architecture.

How many evidence grades does Mira use?

Four: A (official pricing / primary filings), B (App Store and Google Play listings), C (named secondary research), and D (triangulated models). Mira prefers A→B→C and labels D-grade presence as directional so it is never presented as audited % market share.

What SLA should founders expect for a first Mira snapshot?

About 2 minutes (~120 seconds) after you paste a product URL: roughly 90%+ faster than a week-long (~40 hour) analyst sprint for the same founder questions. Support: support@joinmira.ai.

Mira

Architecture

How Mira is built

Blueprint of the product stack, website, Render API, InsForge data, and You.com skills that power live competitive intelligence.

View blueprint

Blueprint architect

System blueprint

Four lanes from the browser to intelligence providers. Skills on the right map 1:1 to chips you see in the product UI.

Mira · v1 blueprint

Edge → control plane → data → intelligence

Lane 01

Edge

Users & delivery

joinmira.ai
Next.js 15 · Vercel
Auth UI
Login · OAuth · sessions
Workspace
Analysis · Labs · Market

Lane 02

Control plane

Render Express API

REST + jobs
Background market / deep dive
Agents
discover · fetch · digests
Cron
24h refresh · weekly mail

Lane 03

Data plane

InsForge

Postgres
Products · competitors · snapshots
Auth
Email · Google · GitHub
Settings
Workspace JSON · keys

Lane 04

Intelligence

You.com skills + Grok

you-web
Search · Evidence · fact-check
you-research
Reviews · talk tracks · overview
you-finance
TAM · investor · market model
you-contents
Pricing pages · site text
xAI Grok
Structure JSON for the UI

Scroll lanes · left to right on desktop

Request blueprints

How a user action becomes a You.com skill call and lands back in the UI with source chips.

Analysis report

1UI /app
2API layers
3you-contents + you-research
4Grok tabs
5Persist analysis-latest

Market model

1UI /market
2Job on Render
3you-finance
4Optional Tavily check
5TAM → SAM → SOM

Evidence locker

1UI Labs
2you-web search
3Grok candidates
4Save to workspace

Deep dive

1UI /company
2Parallel sections
3you-research + you-finance
4Dossier + sources

Hosting split

Vercel

Next.js frontend · joinmira.ai · OAuth routes

Render

cia-api Express · long jobs · cron · secrets

InsForge

Auth + Postgres + storage · always remote

System design

The website talks to a persistent Express API. That API owns long jobs and secrets; InsForge holds data and identity; You.com and Grok supply live research and structure.

joinmira.ai

Browser

Next.js App Router UI: marketing, auth, and the signed-in workspace.

Render · Express

API

Persistent Node process for discovery, market jobs, cron refresh, and digests.

InsForge

Data & auth

PostgreSQL, email/password + OAuth, workspace settings, and storage.

You.com · xAI Grok

Intelligence

Live research and page contents from You.com; structured analysis via Grok.

Request path

Website

Next.js on Vercel, Discover, Market, Company, Distribution

API base

NEXT_PUBLIC_API_BASE → Render Express (Bearer + workspace)

Orchestration

Agents & routes: discover → refresh → market intel → digests

Providers

You.com research/contents/finance · xAI Grok · optional Tavily

Persistence

InsForge Postgres stores products, competitors, snapshots, jobs

Tech stack

What each layer runs, and why it sits where it does.

Frontend

Next.js 15 · App Router, SSR routes, OAuth callbacks
React 18 · Client components for workspace UIs
Tailwind CSS 3.4 · Design system tokens (ink + accent)
Framer Motion / Recharts · Landing motion and in-app charts

API

Express · REST orchestration and long-running jobs
node-cron · 24h competitor refresh + weekly digests
@insforge/sdk · Server-side Postgres CRUD

AI & research

You.com API · research, contents, finance_research
xAI Grok · Structured extraction and analysis (OpenAI SDK)
Tavily (optional) · Independent fact-check for market models

Platform

InsForge · Auth, database, storage
Render · Persistent API (and optional Next host)
Vercel · Preferred frontend host for joinmira.ai
Resend / Zendesk · Transactional email and in-app support

How we use You.com

You.com is the live research backbone. The website never calls You.com directly - the Render API owns the key, rate-limits requests, and maps results into Mira features.

Client module

server/services/youcom.js

Serializes outbound calls with spacing (YOUCOM_MIN_INTERVAL_MS), retries on 429/5xx, and exposes three operations used across agents and routes.

POST /v1/search

Cheap web search for Evidence research and fact-check before finance.

research

POST /v1/research

Competitor discovery, reviews, war-room talk tracks, company overview.

contents

POST /v1/contents

Fetch pricing pages and site text for snapshots and ICP/pricing layers.

finance_research

POST /v1/finance_research

Market sizing, investor enrich, market model, deep-dive financials.

Wired into the product

Each flow starts in the Next.js UI and lands in a You.com call on the API.

Discover competitors

1.UI calls /api/discover
2.API fetches product page via contents
3.research finds rival set
4.Grok structures candidates for approval

Refresh snapshots

1.Cron or manual refresh
2.fetchAgent pulls pages via contents
3.Diff detects pricing / messaging changes
4.Workspace surfaces alerts

Market model

1.Market intelligence job starts
2.finance_research sizes the market
3.Grok builds TAM→SOM narrative
4.Optional Tavily fact-check (You.com fallback)

Company deep dive

1.Company routes request overview
2.research + finance_research enrich
3.Reviews and financials land in UI
4.Distribution uses share / traffic signals

Call sites include discoveryAgent, fetchAgent, routes/intelligence, routes/company, and syndicatedShare. Grok always sits after research when the UI needs structured JSON.

Render, Vercel & why the split

Market jobs and cron need a process that stays alive. That is why the API lives on Render even when the website is on Vercel.

Render

API required

•cia-api, Express server with health check at /api/health
•Holds You.com, xAI, Resend, and cron secrets
•Needed for 2–3 minute market jobs and scheduled refresh
•Optional cia-web can host Next.js from the same blueprint

Vercel

Frontend

•Hosts the Next.js client (root directory: client)
•Serves joinmira.ai / www.joinmira.ai
•Runs OAuth route handlers and feature-request APIs
•NEXT_PUBLIC_API_BASE points at the Render API URL

InsForge

Data plane

•Always remote, Postgres + Auth + Storage
•Used by both the Express API and Next auth proxies
•Workspace isolation and session cookies (cia_auth)

Deploy config

render.yaml defines cia-api (repo root, npm start) and optional cia-web (client build + next start). Production frontend typically uses Vercel with NEXT_PUBLIC_API_BASE set to the live Render API URL (for example *.onrender.com). See the system blueprint for the full lane diagram.

Other integrations

Supporting services around research, messaging, and delivery.

Service	Role	Where
InsForge	Auth, PostgreSQL, storage	server/db · client auth routes
You.com	Live web research and page contents	server/services/youcom.js
xAI Grok	Structured AI analysis	server/services/ai.js
Resend	Digests, invites, support mail	server/services/email.js
Tavily	Market-model fact-check	server/services/tavily.js
Zendesk	In-app messaging widget	ZendeskWidget.jsx
Web Push	Browser notifications (VAPID)	server/services/push.js

See it in the product

Discover competitors, build a market model, and watch You.com-backed research land in your workspace.

Create account Feature requests

Mira architecture

How Mira is built

System blueprint

Edge

Control plane

Data plane

Intelligence

Request blueprints

Analysis report

Market model

Evidence locker

Deep dive

System design

Browser

API

Data & auth

Intelligence

Tech stack

Frontend

API

AI & research

Platform

How we use You.com

Wired into the product

Discover competitors

Refresh snapshots

Market model

Company deep dive

Render, Vercel & why the split

Render

Vercel

InsForge

Deploy config

Other integrations

See it in the product