For school psychologists & BCBAs

Observe.
Analyze.
Write-up.

Keyboard-driven behavioral observation with automatic cross-session analysis and an AI-assisted write-up workspace. First session to final report section — one tool.

Request beta access → Free during beta · No account required
Session: J. Martinez · Math Block · Active ●
Behaviors
Out of Seat O
Vocalization V
Aggression A
ABC Tags
Demand Escape Peer Attn
14:02:11
Live Observation Feed
14:01:45Duration Started
Out of Seat
A: Independent Work
13:58:20Frequency +1
Vocalization
Active Mode
Heads-up Keyboard Coding

One tool from session to write-up.

Stop tallying on paper, re-entering into a spreadsheet, and reformatting for the report. Sightline connects every phase of the observation workflow.

01
Observe

Eyes on the student, not the screen.

Hit space to start. Score behaviors with number keys. Log ABC events with Tab. The timer, intervals, and scoring all run together, keyboard-driven so you never look away from the student.

02
Analyze

Patterns surface. No spreadsheet required.

Graphs update as sessions are added. Filter results by activity phase to see where behavior concentrates. Antecedent-consequence chains and cross-session trends appear automatically — no manual compilation.

03
Write

Your data is already in the draft.

The write-up workspace puts session evidence alongside your narrative. AI drafts from your coded data — every claim traceable to a specific interval or session. Review, revise, and drop it into your report.

Optional · On-Device · Private

AI that accelerates the write-up.

Everything in Sightline works without AI. But when you're ready to draft, targeted AI features turn your coded data into report-ready language — running locally on your machine, with every claim traceable to a specific session.

Narrative Drafts

Clinical-language observation summaries with temporal grounding and phase-aware comparisons. Start from a grounded draft instead of a blank page.

Behavioral Portraits

Cross-session synthesis that identifies consistent patterns, activity-specific concentrations, and changes over time across your observations.

Function Hypotheses

Conditional probabilities from ABC data with suggested behavioral functions — escape, attention, tangible, sensory — grounded in your coded events.

Session Digests

One-paragraph summaries with key statistics and notable patterns. Scan your caseload at a glance.

Example: AI-generated narrative excerpt

During a 10-minute partial interval observation, off-task behavior was recorded in 60% of intervals (12 of 20). Off-task behavior was most concentrated during the first half of the session, with occurrences in 8 of the first 10 intervals. Following a verbal redirect at interval 6, a brief reduction was observed, though the behavior resumed within two intervals.

Architecture: Local-First

Legally defensible.
Locally processed.

Student data shouldn't be parsed by a startup's cloud server. Sightline runs entirely on your device. No cloud vendor, no BAA negotiation, no breach surface.

💾

No Cloud Vendor

Your database lives on your hard drive. Zero external breach risk. No IT approval process or data agreement required.

🧠

On-Device AI

AI narrative drafting runs locally via Ollama. No API calls to external services. Optionally connect your org's existing cloud AI agreement.

🤝

Zero BAA Negotiation

We never transmit or hold student data, so complex district Business Associate Agreements are entirely sidestepped.

Privacy architecture →

Built for how observation actually works.

Rigorous enough for an FBA. Intuitive enough for your first direct observation.

Eight recording methods.

Every standard recording type — from partial interval to ABC to latency — available when you need it. Pick the method that fits the question you're trying to answer.

ABC Recording Duration Partial Interval Whole Interval Momentary Time Sampling Frequency Latency Narrative

Ready-to-use protocols.

Pre-configured templates for common evaluation scenarios.

On-task / Off-task
Disruptive behavior
Academic engagement
A B C

Built for FBAs.

Fluid ABC tagging with number keys. Antecedent-consequence chains, conditional probability tables, and function hypotheses computed from your coded events.

Customizable per student.

Behavior definitions, keyboard bindings, quick phrases, and ABC tag sets live at the student level. Global defaults, student-specific overrides on top.

See all features →

Start your next observation here.

Sightline is in closed beta for macOS and Linux. Free during beta, no account required.

Request beta access →

No data leaves your machine