Everything Humane AI does today
Understand each user over time, keep them safe, bring a person in when it matters, and respond at a human pace, on a platform built for developers.
Understand users
User state across sessions
Humane keeps a running state for every end user, from mood and energy to trust and grudge, each with its own behaviour over time. Your model finally knows whether it is talking to someone on a good day or at the end of a bad week.
Explore User state across sessionsObserve mode
observe() reads a turn your own model already produced and returns the user's state and a risk level. It needs no model configuration, generates nothing and never sits in front of your reply.
Explore Observe modeAnomaly detection
A scheduled sweep compares the last day with a 30-day baseline for each user and each tenant. When a metric moves two or more standard deviations, Humane records an anomaly, rates its severity and fires anomaly.detected.
Explore Anomaly detectionDurable memory
Every turn is stored in a per-tenant vector memory, searchable by meaning. Seed it from your other systems, search it from your own code and erase it when a user asks to be forgotten.
Explore Durable memoryJourney and history
The journey endpoint returns everything that concerned one user in time order, from their messages to safety events, proactive triggers and experiment assignments. Conversation history can be read or cleared per user.
Explore Journey and historyKeep users safe
Safety gates
Before a reply reaches the user, Humane runs it through a fixed stack of gates and returns PROCEED, HOLD or BLOCK. Each gate records why it decided, and a gate that fails resolves to HOLD, never to PROCEED.
Explore Safety gatesRisk alerts
Elevated or critical risk fires a signed safety.flagged webhook from both observe() and process(), whether or not a gate also acted. Send it to your own endpoint, Slack or Discord, and on to a human takeover.
Explore Risk alertsPII scrubbing
Humane scans each outbound reply for personal identifiers. Phone numbers, emails, addresses and similar are masked in place with typed markers, while SSNs, card numbers and IBANs block the reply entirely.
Explore PII scrubbingPolicy rules
Policy rules match on a user's state, their message or the time since they were last seen, then fire an event, adjust settings for that turn, set the safety action or skip the model. Rules are validated when they compile and can be tested before they go live.
Explore Policy rulesDependence safeguards
The emotional dependence gate flags replies that promise permanence, exclusive availability or a unique understanding of the user. When a user's conversation trips it three times within an hour, Humane slows replies and shifts to a boundary-respectful tone until the window passes.
Explore Dependence safeguardsHuman in the loop
Human takeover
When the signals say someone needs a person, an operator opens a takeover for that user. The AI pauses, the operator's messages reach the user in the same conversation, and the AI resumes when the operator releases it or goes idle.
Explore Human takeoverOperator drafts
The drafts endpoint asks your configured model for a few short reply options, built from the recent conversation and the user's current state. Operators can steer them with a line of guidance, then use, edit or ignore them.
Explore Operator draftsProactive triggers
Humane watches each user's state and fires an event when a threshold is crossed: mood dropping, energy running low, a user going quiet, trust passing a milestone. Each trigger has a cooldown and carries a suggested action, and your app decides how to reach out.
Explore Proactive triggersTask detection and escalation
When a user writes “remind me to call the pharmacy at 6pm”, Humane extracts the task and its time without a separate API call. Open tasks climb a five-level ladder, firing reminder and escalation events until someone completes or snoozes them.
Explore Task detection and escalationRespond well
Timing engine
Humane recommends how long to wait before each reply, based on the user's mood, energy, the intensity of their message and the pace of the conversation. Every delay is broken into labelled contributors, so you can see and chart why.
Explore Timing engineTone and context
Before your model runs, Humane assembles its prompt from the user's state, relationship, recent conversation, relevant memories and open tasks, scoring and trimming each section to fit a budget. It also sets the tone, empathy, formality and length the reply should take.
Explore Tone and contextPlatform
Model adapters
Humane talks to models through four adapters: OpenAI-compatible, Anthropic, Amazon Bedrock and Replicate. Change providers without touching the behavioural layer.
Explore Model adaptersStreaming
stream() returns server-sent events: a metadata event with the user's state, safety decision and timing before the first token, then the tokens, then done. If a gate blocks, the stream says so and ends.
Explore StreamingWebhooks and events
Humane posts events to your endpoints as things happen: safety flags, holds and blocks, proactive triggers, takeovers, anomalies and more. Every delivery is signed with HMAC-SHA256, retried on failure and parked in a dead-letter queue if it keeps failing.
Explore Webhooks and eventsTelegram and WhatsApp
Register a Telegram bot or a WhatsApp Business number and Humane receives messages, runs them through the full pipeline and delivers the replies. Inbound webhooks are verified, and outbound replies are split, paced and tracked.
Explore Telegram and WhatsAppExperiments
Define an experiment with two variants, each a set of engine settings, and Humane assigns users and applies the right overlay on every call. Results compare the variants with lift and a two-proportion z-test.
Explore ExperimentsAnalytics and cohorts
The Behavior Score is a 0–100 composite of retention, conversation depth, response to proactive outreach, abandonment and sentiment trend. Cohort views show retention and average mood by month of first contact.
Explore Analytics and cohortsTeams, SSO and roles
Each team's data is isolated at the query layer and gets its own memory store. Members sign in with email or Google, hold roles that control what they can change, and manage API keys that can be created, rotated and revoked.
Explore Teams, SSO and rolesPlayground
The playground runs Humane's analysis on a message and shows what it detected and how it would shape the reply. Point it at a model and compare the raw reply with the Humane-shaped one side by side.
Explore PlaygroundAudit trail
Humane records gate outcomes, safety flags, webhook deliveries, engine decisions and admin actions, each with its reason. When someone asks why the AI did something, you can show them.
Explore Audit trailEvery AI should know who it's talking to.
The safety and human-oversight layer for emotional AI: user state across sessions, risk alerts, safety gates and human takeover, with every decision explained.
