StoryCoach K12
The pen stays in the student's hand.
StoryCoach reads a student's draft and answers with questions and highlighted passages, never with rewritten sentences. Grades 3 through 12, one coaching conversation at a time, with a separate safety check reading alongside every session.
Contact us
Watch
See StoryCoach at work
The coach answers beside the draft, never inside it.
Last spring our class planted a garden behind the gym. We grew tomatoes, beans, and a sunflower that got taller than the fence. It was really good and everyone liked it.
You wrote:
“It was really good and everyone liked it.”
What did you mean when you said “really good”? Can you tell me more about that?
Things to try
- 1Show one moment in the garden instead of telling us it was good.
- 2Who is everyone? Name one person and what they did.
- 3The sunflower taller than the fence is your strongest picture. Try ending with it.
Formative score 6 / 10
Every session is a conversation, not a correction.
Draft and ask
Students sign in with a class code, a first name, and a PIN, no email address required, then open the writing editor. The draft panel is theirs alone, a locked editing pane no coaching process can enter, and it saves automatically every few seconds. When there is at least a sentence or two on the page, the student clicks “Get feedback.”
Respond
The coach answers in a separate panel, never inside the draft itself. It quotes the student’s own words before commenting, highlights the passages worth a second look, and asks questions, “What did you mean when you said…?”, “Can you tell me more about that?”, rather than supplying better sentences. One to three suggestions arrive alongside a formative score out of ten, calibrated to grade band and writing mode.
Decide
Suggestions are just suggestions. The student can revise, ask again, or leave a passage exactly as written. Most students do their best work after two to three rounds, and nothing in the exchange requires them to accept a single word the coach offers.
Compound
Every round quietly feeds the record. The score, the standard practiced, and the full transcript move into the teacher’s view of standards mastery, the same signal that later surfaces students who may benefit from extra support. Coaching accumulates. The draft still belongs to the student who wrote it.
Enforceable commitments, not aspirations.
Every row below is a claim a district's legal or IT team can verify, not a promise offered on faith.
Commitment
Enforcement
The student's draft text is never modified, replaced, or hidden by any AI process.
Enforcement: Hard-coded, enforced at the API level. Not a policy setting.
The student never sees the safety-flag process.
Enforcement: Enforced in the UI architecture itself.
Every safety-flag action a teacher takes is logged.
Enforcement: Written to a tamper-evident audit log with timestamp, teacher ID, and rationale.
No automated escalation ever leaves the school.
Enforcement: No such pathway exists in the system. A flag only ever reaches a teacher's queue.
No student personal information enters a coaching prompt.
Enforcement: Enforced. Prompts contain draft text and a grade-band tag only, nothing else.
Student writing is never used to train outside models.
Enforcement: Written into our data-processing agreement with every district, not a marketing promise.
A teacher must own the class before a student can access it.
Enforcement: Enforced by authorization middleware on every session.
Baseline security: rate-limited logins, traffic encrypted in transit.
Enforcement: No plain-HTTP fallback, configurable per district.
Safety and wellbeing
What the safety check actually does
A separate classifier reads roughly the last 2,000 characters of every draft in parallel with the coaching call. It never delays the student's response, and it never appears anywhere in the student's experience.
Categories
- 01Self-harm
- 02Abuse disclosure
- 03Violence threat
- 04Severe distress
- 05Substance crisis
- 06Other sensitive content
Severity
What it means
What happens
None
Ordinary writing, including age-appropriate dark themes.
Logged silently. No alert.
Low
A sensitive theme that reads as fictional or emotionally distant.
Logged silently for analysis. No alert.
Moderate
First-person framing, specific detail. An adult should read this soon.
Appears in the teacher's Concerns dashboard.
Severe
Explicit, immediate, first-person language about harm.
Appears prominently. The teacher sees it on next login.
Teachers see
A private Concerns tab: the evidence quote, category, severity, date, and a link to the full session transcript.
Parents see
A read-only sensitive-content notice when a flag is open for their child, no details, just that the teacher has been alerted and is reviewing.
Students see
Nothing. Their coaching experience is unchanged whether or not a flag fires.
Deliberately out of scope, by design
- Automated parent notification when a flag fires
- Counselor handoff or case management workflow
- Student-facing crisis resource panels
- Cross-district analytics on flag rates
These are not technical limitations. They are deliberate policy boundaries, held pending formal consultation with school counselors, district policy owners, and mandatory-reporter compliance experts.
Five roles, one shared record
Students
Sign in with a class code, a first name, and a PIN, no email required. Write, ask for feedback, and decide what to revise.
Teachers
Build classes, assign writing modes, monitor every transcript, resolve Concerns flags, and review AI-drafted plans before activating them.
Parents
Read-only, linked with a family code. See scores, download finished stories, and get a sensitive-content notice when a flag is open, nothing more.
Principals
A building-level view of Targeted Practice tiers and open safety concerns. Can view any support plan but cannot activate or close one.
District
Schools, teachers, and the audit log across the district. Can drill into any teacher's roster read-only, and add or remove staff accounts.
ATLAS rubric scoring
Beyond the quick formative score, a teacher can apply the ATLAS analytic rubric to any session for a structured evaluation across four dimensions, Purpose & Focus, Organization, Evidence & Elaboration, and Conventions, each scored zero to four.
4 dimensions, 16 points total
ATLAS is teacher-only. Students never see it, and scoring never touches the draft.
Targeted Practice
Nine writing-growth indicators run quietly in the background. A student four or more weeks below grade-level expectation, or with two or more risk indicators, surfaces in the teacher's “Students who may need support” panel.
From there a teacher can request an AI-drafted Tier 2 Early Support Plan, goals, strategies, and monitoring checkpoints included. The AI writes the first draft; the teacher reviews, edits, and activates every section. Only a teacher can move a student to Tier 3.
What StoryCoach promises, in writing
FERPA-aligned access
Student records are visible only to the assigned teacher, school administrators, and linked parents. No cross-school or cross-district visibility without explicit permission.
Class-code login
Students sign in with a class code, a first name, and a PIN. No student email address, ever.
No data resale
Student writing and performance data is never sold, licensed, or used to train outside models.
No behavioral advertising
No device fingerprinting, no advertising profile, no social graph. The system collects only what a coaching session requires.
Questions districts ask first.
Talk with us about StoryCoach for your district.
We'll walk through a real session, the safety architecture, and how the record feeds Targeted Practice for your teachers.
Contact us