llms.txt
# Sorcrr
> Human-centered sourcing, transparent evidence, and permissioned assistance.
Sorcrr is a multi-product operating platform built from 15+ years of real-world hiring and systems work. It combines bounty-driven referrals, structured video, collaborative workflows, a versioned Career Rank Score, and SAI while keeping consequential decisions with people.
## Products
1. **Hiring** ([sorcrr.com](https://sorcrr.com)) — Video-first recruiting, collaborative evaluation, transparent career evidence, referral bounties, and human-led decisions
2. **Market** ([market.sorcrr.com](https://market.sorcrr.com)) — Service marketplace across 17 verticals: 15 public direct-checkout verticals, Healthcare in provider preview, and Startup/SaaS redirect attribution
3. **Books** ([books.sorcrr.com](https://books.sorcrr.com)) — Personal and company operating books that keep native, Hiring, Market, invoice, refund, and reversal events in their correct owner scope
4. **SAI** — One permissioned assistant layer that can route requests across authorized Hiring, Market, Books, company-knowledge, and customer workflows
## Hiring
- [How Sorcrr works](https://sorcrr.com): Create a video job description, fund the disclosed posting economics, share through the network, review candidate evidence, collaborate and interview, then hire and reconcile
- [Career Rank Score and Open Algorithm](https://sorcrr.com/crs): Canonical additive career-evidence score with no ceiling, plus deterministic viewer-owned filtering and sorting
- [How SAI works](https://sorcrr.com/sai): Permissioned context, deterministic and model-assisted routes, AI-credit boundaries, and human controls
- [Pricing](https://sorcrr.com/pricing): Hiring and Market economics
- [About](https://sorcrr.com/about): Origin story, values, and product system
### Career Rank Score (CRS)
CRS means **Career Rank Score**. The current `crs-v5` formula is deterministic, versioned, reproducible, and viewer-independent. Canonical CRS = profile foundation + the sum of every eligible review:
- Dated work experience: up to 150 points, reached at 15 merged years (overlapping roles never count twice)
- Completed education: up to 90 points
- Skills and certificates: up to 60 points
- Each eligible submitted four-dimensional review adds its raw Rating + Integrity + Intelligence + Energy (each on the native 0–5 scale), up to 20 points per review; private and anonymous complete reviews still count
- Unbounded above, with no renormalization: a missing component adds nothing rather than being scaled up
- No multiplier for relationship, reviewer CRS, verification, recency, account tier, employer, or institution
- Until the first eligible review, the foundation appears as a faded potential CRS
- Eligible review count, distinct reviewers, verification coverage, formula version, freshness, availability, and component breakdown are shown beside the score
- Historical v4 scores remain readable in score history and are never relabeled
- No age, account age, salary, referral count, or employer/institution prestige is used
CRS never makes an automatic employment decision. No eligible review means unrated, not a negative score.
### Open Algorithm / Build Your Own Algorithm
Open Algorithm is a deterministic ranking lens, not AI and not a replacement CRS. Authorized users can save personal, company, or job-scoped policies using:
- All eligible or verified-only evidence views (verified = reviews backed by a paid Market booking), plus an optional exclude for reviews by a stated reviewer relationship, such as friends
- Canonical score, explicit weighted dimensions/components, or score per verified year
- Score, confidence, verification-coverage, and verified-work-year thresholds
- Immutable policy versions and server-side authorization
Every result retains the original CRS and returns the derived value, confidence, policy version, and explanation. Protected traits are not valid filters or weights. Humans keep final control.
### SAI
SAI is one clearly identified AI assistant. The interface routes by intent and permission rather than asking users to understand technical agents or personas.
- Personal, company, customer-facing, and employer-only contexts remain separate authorization boundaries
- SAI separates live session context, user-scoped memory, approved knowledge, and derived retrieval indexes rather than treating an account as one unrestricted context pool
- Supported AI settings expose memory inventory and source drill-down, retention preferences, compaction previews, and bounded cleanup previews; canonical source records keep their own retention and permissions
- Deterministic tools—such as applying or explaining a CRS ranking receipt—do not invoke a language model or consume generative-AI credits
- Drafting, summarization, or interpretation may use a configured enterprise model and AI credits
- Where available, a user can disable eligible profile-context personalization
- Important external, account, financial, customer, booking, and hiring actions retain normal permissions and confirmation
- Model output is assistive and may be wrong; users must review important output
- Interview intelligence is separately gated and is not implied to be active in every interview
## Referral Bounties
Hiring and eligible Market direct-checkout flows can fund referral bounties. The server computes the eligible chain and uses the shared edge-favored geometric distribution with a maximum depth of five. Sharing or appearing in a chain does not guarantee payment.
## Sorcrr Market
Market has 15 public direct-checkout verticals. Healthcare is a provider preview: public discovery, booking, checkout, and referral rewards remain closed. Startup/SaaS uses redirect attribution, with rewards only under explicit signed third-party terms.
Providers own their brand, pricing, customer relationships, credentials, service delivery, taxes, and applicable professional obligations. Market ratings and verified-review evidence remain separate from the hiring-only CRS.
## Legal and Privacy
- [Privacy Policy](https://sorcrr.com/privacy): CRS, Open Algorithm, SAI context, provider, retention, and user-rights disclosures
- [Terms of Service](https://sorcrr.com/terms): Human-review, lawful-use, employer notice/audit, correction, and automated-employment-tool responsibilities
- CRS, Open Algorithm, SAI, matching, and summaries are decision support—not autonomous hiring, firing, booking, financial, or professional decisions
## Contact
- Email: hello@sorcrr.com
- X: @Sorcrrs
## More Detail
- [llms-full.txt](https://sorcrr.com/llms-full.txt): Extended platform and feature details