Mirror, the genome-aware AI agent

Mirror is the AI agent at the centre of Haeckel. Every other surface in the platform is a structured view of one slice of your data, while Mirror is the place where you can ask any question that crosses those slices. You can talk to Mirror by voice or by typing, and Mirror responds aloud while drawing visual artifacts on the canvas to illustrate its answer.

The model stack

Mirror is not a single model. The chat surface runs on a frontier conversational model from a top-tier provider, chosen because the cost of a wrong genomic explanation is high enough that the marginal expense of a better model is worth paying. Two cheaper models handle the auxiliary tasks: a mid-tier model parses natural-language queries into structured search criteria for the Networks AI Candidate Finder, and a fast classifier runs moderation and intent detection where the latency budget matters more than headline quality.

What Mirror sees in its context window

Each turn assembles a structured context that the model receives as part of its prompt. The context contains, in order:

1. A system prompt that fixes Mirror's identity, its scope, and its non-clinical posture (it never prescribes).
2. A digest of your analytical results across all seventeen pipeline modules, with the most relevant slices brought to the top based on the conversation so far.
3. Your network memberships, the artifacts you have starred, the people you have connected with, and any flags you have set on your privacy preferences.
4. A running summary of your prior conversations across sessions, kept compact through periodic compression so the context window never overflows.
5. The current message and a short look-back of the last few turns in this session.

Tool use: Mirror can do things, not just describe them

Mirror is wired to the platform with a small set of tools the model can invoke during a turn. When you ask "how does my CYP2D6 status interact with antidepressants", the model can call a tool that fetches the drug-gene edges from the Deterministic Biological Knowledge Graph rather than guess from training data. When you ask "compare my ancestry to the cohort average", the model can call a tool that computes the cohort statistics live. When you ask "share this artifact with Eve", the model can call a tool that creates the share. The tools surface explicitly in the conversation transcript so you can audit what the model did on your behalf.

Memory across sessions

Mirror persists a running summary of your prior conversations, keyed by your user ID. The summary is generated after each session, distilling the salient facts (topics covered, decisions you made, follow-ups you flagged) into a compact passage. On your next visit, that summary is prepended to the context window so Mirror does not need to be reintroduced. You can review or clear the memory from Settings at any time, and clearing it is immediate.

How your data reaches the model

Mirror conversations go to a frontier-model provider that operates under a no-training policy for API traffic, so your prompts and the model's responses are not used to train future models. Voice synthesis goes to a separate provider, again under a no-training contract. The platform never sends your raw DNA file to any third party. What leaves Haeckel's servers, per turn, is the assembled context (digest plus history) plus your current message; what comes back is the model's response and any tool calls.

Embeddings used by the Networks recommender go through a separate embeddings provider, also under a no-training contract. Every third-party that touches user data is listed in the privacy article under Subprocessors.

Latency

Mirror is designed to feel conversational rather than chatty. The first token of a response arrives quickly enough that you can start reading before the full answer has finished streaming, and voice synthesis starts as soon as the first sentence completes so you are not waiting for the whole reply before Mirror starts speaking. Most users settle into asking complete questions rather than firing off short prompts, because the model rewards depth.

What Mirror does not do

• Prescribe. Mirror will explain what an APOE ε4 carrier status implies for Alzheimer's risk, but it will not tell you to start any specific medication or supplement.
• Diagnose. A polygenic score above the 95th percentile is not a diagnosis, and Mirror frames it as an inherited disposition rather than a present condition.
• Improvise outside its competence. When asked about something genuinely outside genomics (general medical advice, psychotherapy, legal questions), Mirror declines and points you to a human professional.
• Forget on demand without confirmation. Clearing memory requires a confirmation step in Settings to prevent accidental data loss.
• Chat across users. Each session is scoped to one authenticated user; Mirror cannot answer questions about another person's genome unless you have explicitly granted access through Networks or sharing.