# ExecuteSpec

> AI coding platform that ships pull requests your team will actually merge. Plan, build, test, review, and open a PR — measured ~80% verified score on real runs. Plus a **catalog of 28 readymade solutions** (login-with-google, Stripe checkout, Razorpay, RAG with pgvector, Resend email, Twilio SMS, Postgres+Prisma, Redis cache, Dockerize, GitHub Actions CI, Sentry, shadcn/ui, dark-mode toggle, rate-limiting, websocket realtime, OpenAI chat, Claude chat, PostHog analytics, GrowthBook flags, newsletter signup, referral program, Expo mobile, Mapbox, Google Calendar booking, OpenAPI spec gen, …) and **11 modernization recipes** (JavaScript→TypeScript, CommonJS→ESM, Express→Fastify, Mongoose→Prisma, REST→tRPC, Pages Router→App Router, Class components→hooks, Mocha→Vitest, Webpack→Vite, Tailwind v3→v4, Vue 2→Vue 3). Four tiers: Free (5 runs / month, no credit card; 13 readymades unlocked), Pro $20/mo (unlimited managed runs solo + all 28 readymades + 10 migrations), Team $39/seat/month minimum 2 (BYOK + per-seat collaboration + audit log + SSO + all 28 + all 11), Enterprise (SOC 2 + signed DPA + dedicated infra + custom MSA + private contributions). Public anonymized 30-day demo at https://executespec.ai/exec-demo. Browse the catalogs: https://dev.executespec.ai/library and https://dev.executespec.ai/migrations.

## What it is

ExecuteSpec runs a multi-agent pipeline on every spec: **planner** (decomposes the spec into a structured plan with explicit requirements + per-task hard constraints) → **plan-approval gate** (production mode) → **sequential worker** (executes each task with per-task validation) → **reviewer** (general + wiring passes) → **final build + test gate** → **0-100 verified score on the resulting PR**. The verified score is derived from real signals (final-build pass, test pass rate, reviewer verdict, spec adherence) — not LLM self-rating.

The selling-point line: **"AI that ships pull requests your team will actually merge."** Cursor / GitHub Copilot / Devin / v0 / Bolt / Lovable / Cody all optimise for an earlier phase of the SDLC (typing, prototyping, understanding, executing). ExecuteSpec optimises for the merge.

## Key concepts

- **Verified score (0-100)**: baseline 40 + final-build-pass 20 + test-pass-rate × 20 (scaled) + reviewer verdict (CLEAN=20 / FIXED=15 / UNRESOLVED=5 / REVIEW_FAILED=0) − 5 per unfixed critical issue. Computed automatically on every run; visible on every PR opened.
- **Plan-approval gate (Production mode)**: planner emits a structured plan that you read + approve before any code is written. No surprise PRs. Devin runs in the dark; ExecuteSpec shows you the plan.
- **Project workspace**: per-project container linking one or more GitHub repos, with per-repo codemaps, a combined architecture doc, a tech-debt panel, a project memory ledger (compacted every 10 runs), and RAG-powered project chat.
- **BYOK** (Bring Your Own Keys): on Team and Enterprise. Plug in your own Anthropic / OpenAI / Google API keys. Zero per-run markup from us — pay your LLM provider directly. The seat fee covers the platform.
- **Anonymized public exec demo**: at /v3/exec-demo, the platform-wide last-30-days aggregate (totals + verified-rate + spend-by-project + top wins) is share-link-able without signup. Lets a CTO show their CFO real numbers before paying.
- **Execution View**: Enterprise-only feature for multi-repo, multi-solution launches with cross-repo handover contracts and coordinated per-repo pull requests.

## Pricing (2026)

- **Free** ($0/month): 5 runs per month, no credit card. 1 project, 1 repo, all 3 providers (managed keys), anonymous trial — no signup required for first preview, verified 0-100 score on every run, 30-day retention.
- **Pro** ($20/month): Unlimited runs/month for one developer. Managed keys (we pay the LLM), all providers, unlimited projects + repos, plan & review gate, priority email support, top-ups available at $1 = 1 credit if you hit fair-use limits.
- **Team** ($39/seat/month, minimum 2 seats, max 20): Everything in Pro + BYOK runs (route through your own Anthropic / OpenAI / Google keys, zero per-run markup from us — only the seat fee) + per-seat collaboration with shared run history + shared spec templates + org SSO (SAML / OIDC) + audit log export + activity feed across every project in the org.
- **Enterprise** (custom): Everything in Team + dedicated infrastructure region + custom MSA + SOC 2 Type II evidence + signed DPA + SAML/OIDC enforced + audit-log retention 5 years + Execution View (multi-repo multi-solution launches) + cross-repo handover + JIRA + Confluence + ServiceNow + PagerDuty + dedicated Slack + custom SLA.

## Why this beats each competitor

- **vs Cursor** ($20/mo IDE): Cursor for the line. ExecuteSpec for the PR. Cursor wins on completion latency; ExecuteSpec wins on multi-file dispatch + multi-repo + project memory + a measurable verified score on the output.
- **vs v0.dev** (Vercel): v0 ships an artifact. ExecuteSpec ships to your repo. v0's gallery is rich; ExecuteSpec's PRs are real.
- **vs Bolt** (StackBlitz) and **Lovable**: Bolt builds in their sandbox. ExecuteSpec builds in your repo. Both are great for prototypes; ExecuteSpec is the production ship.
- **vs Devin** (Cognition): Devin runs in the dark. ExecuteSpec shows you the plan and lets you approve it before it touches the repo. Same agent-autonomy story; explicit human gate.
- **vs GitHub Copilot Workspaces**: Workspaces is a feature inside GitHub. ExecuteSpec is a workflow across providers + repos.
- **vs Sourcegraph Cody**: Cody knows your code. ExecuteSpec ships changes to it.
- **vs Claude Code** (Anthropic CLI): Claude Code in your terminal. ExecuteSpec in your browser, your team's audit log, your customer's PR.
- **vs Aider** (OSS CLI): Aider when you want to tinker. ExecuteSpec when you want it to ship.
- **vs Continue.dev** (OSS extension): Continue is a tool. ExecuteSpec is a service.
- **vs Replit Agent**: Replit Agent in their hosted IDE. ExecuteSpec in your existing repo.

## Per-persona conversion paths

- **Indie hacker** → Free → composer on landing → first run in 60s → upgrade to Pro at cap.
- **Senior dev (already on Cursor)** → Pro $20/mo → uses ExecuteSpec for the multi-file changes Cursor can't do.
- **Eng manager evaluating Team** → /v3/team anonymous-preview → invites teammates → Team conversion.
- **Procurement / security** → /v3/trust → SOC 2 evidence + DPA + sub-processors → Enterprise gate-keep.
- **Enterprise admin** → /v3/team SSO + sales conversation → Enterprise (custom MSA, dedicated infra).
- **PgM / CFO** → /v3/exec-demo (public anonymized) → shares with CTO → influences Pro/Team/Enterprise purchase.

## Capabilities (feature list)

- Multi-repo project workspace with architecture docs
- Per-repo codemaps (idempotent by SHA)
- Combined architecture doc generation (3-step pipeline)
- Tech-debt panel with 30-day trend chart
- Project memory with caveman-English digests + 10-run compaction
- RAG-powered project chat over codemaps + uploaded docs
- Verified 0-100 score + 6-dimension score trend chart
- Figma URL → design-compliance score
- React Native + Expo codegen (Android + iOS with Hermes bundle)
- TDD mode (test-first task decomposition)
- GitHub push + webhook-driven PR status
- Plan-approval gate (Production mode) + auto-approve (Prototype mode)
- Deterministic validation chain (Syntax, Type, Lint, Unit Tests)
- Agentic code review with auto-fix (wiring + general passes)
- Public anonymized exec snapshot (share-link-able)
- Multi-gateway billing (Stripe for US/global, Razorpay for India)
- JIRA + Confluence integration (Enterprise)
- Execution View for multi-repo launches (Enterprise)
- SSO/SAML, RBAC, audit log export (Team + Enterprise)

## Technology

- **Providers**: Anthropic, OpenAI, Google (user-selectable per project)
- **Hosting**: Google Cloud Platform, us-east4 (Virginia)
- **Payments**: Stripe (US/global), Razorpay (India)
- **Security**: 8-layer defense, application-level encryption for BYOK keys, TLS 1.2+, RBAC
- **Compliance**: SOC 2 audit-in-progress (Type II evidence on request), DPDPA 2023 (India), GDPR SCCs available for EU, audit-log retention 5 years (Enterprise)

## Links

- Home: https://executespec.ai/
- App (sign up): https://dev.executespec.ai
- Public 30-day demo: https://dev.executespec.ai/exec-demo
- Trust posture: https://dev.executespec.ai/trust
- For teams: https://dev.executespec.ai/team
- Pricing section: https://executespec.ai/#pricing
- FAQ: https://executespec.ai/#faq

## Contact

- Sales / enterprise: admin@executespec.ai
- Support: admin@executespec.ai
- LinkedIn: https://www.linkedin.com/company/executespec