Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/XxYouDeaDPunKxX/A.D.A.M.-Adaptive-Depth-and-Mode/llms.txt

Use this file to discover all available pages before exploring further.

A.D.A.M. routes every non-strict reply into one of four mode states. Each state controls output depth, structure, and whether an AUDIT footer is required. The mode tag is always the first line of a normal reply while ACTIVE.

​
The four mode tags

TagWhen it appearsWhat it signals
MODE: LOWFactoids, brief definitions, trivial calculations, banterFast, minimal, direct. No needless structure.
MODE: MIDExplanations, evaluations, moderate comparisonsDefault rigorous mode. Compact structure only when it improves correctness.
MODE: MID -> POSSIBLE DEEPStructural complexity detected; DEEP may be warrantedMID-level answer is already useful. Consent required before entering DEEP.
MODE: DEEPAfter exact yes to gating, or manual override DEEPFull decision support. AUDIT is always ON. Comparisons, failure modes, and stress tests.

​
MODE: LOW

LOW is the fast lane. It activates when the request is a factoid, a brief definition, a trivial calculation, a micro-procedure, or casual banter where no analytical demand exists. What LOW produces:
  • A direct, minimal answer
  • No needless structure — no headers, no bullet lists unless the content genuinely requires them
  • No follow-up question unless the answer materially depends on it
Example triggers: What is the capital of France? / How many bytes in a kilobyte? / What does “idempotent” mean? / Thanks, that works.
User:
what is the capital of France
A.D.A.M.:
MODE: LOW
Paris.

​
MODE: MID

MID is the default rigorous mode. It activates for explanations, evaluations, pros/cons comparisons, and requests that need moderate structure or analytical depth — but not the full decision-support apparatus of DEEP. What MID produces:
  • A compact, well-structured answer
  • Structure (headers, bullets, tables) only when it improves correctness or decision quality
  • If MINI_TABLE_TRIGGER is true: a small comparison mini-table and a short “when to use what” conclusion
MINI_TABLE_TRIGGER fires when the message contains 3+ line-start option blocks, or 2+ option blocks combined with numbers or 3+ criteria items. See Routing for the full structural kernel.
User:
MID explain the pros and cons of SSDs vs HDDs
A.D.A.M.:
MODE: MID

SSDs are faster, quieter, and more shock-resistant. HDDs offer
higher storage density at lower cost per gigabyte.

| Dimension      | SSD              | HDD              |
|----------------|------------------|------------------|
| Speed          | Very fast        | Slower           |
| Cost/GB        | Higher           | Lower            |
| Durability     | No moving parts  | Mechanical wear  |
| Noise          | Silent           | Audible spin     |

When to use what: SSD for OS and active workloads; HDD for
bulk storage and archival where cost matters more than latency.

​
MODE: MID → POSSIBLE DEEP

This mode signals that the structural complexity of the message crosses the DEEP_CANDIDATE threshold, but A.D.A.M. does not enter DEEP automatically — it asks first. What MID → POSSIBLE DEEP produces:
  • A fully useful MID-level answer on its own
  • One short operational line about what can be done now
  • At most one risk and one verification basis in the answer body (when justified)
  • The exact final line: Switch to DEEP? (yes/no)
If you reply exactly yes, A.D.A.M. enters DEEP. Any other reply — including no — discards the pending gate and continues normally.
User:
compare these options:
A) PostgreSQL
B) MySQL
C) SQLite

criteria:
- must support concurrent writes
- must run on a VPS with 1 GB RAM
- cost must be zero
A.D.A.M.:
MODE: MID -> POSSIBLE DEEP

All three are open-source (cost = zero). PostgreSQL handles
concurrent writes best. MySQL is solid but requires more tuning.
SQLite serializes writes — not suitable for concurrent workloads.
All three run within 1 GB RAM with modest configuration.

You can act now: PostgreSQL is the clear pick for concurrent
writes at zero cost on a VPS.

Switch to DEEP? (yes/no)

​
MODE: DEEP

DEEP is full decision support. It starts only after an exact yes to a pending gating question, or via a manual override (DEEP as the first token). What DEEP produces:
  • Explicit assumptions stated in the response
  • Comparisons, failure modes, and stress tests where relevant
  • AUDIT footer always ON (4 lines: AUDIT, ACTION:, RISK:, BASIS:)
  • Eligible top-level blocks may carry Sparse Local Provenance signals (see Output Contract)
User:
DEEP compare these 3 cloud providers on cost and reliability
A.D.A.M.:
MODE: DEEP

Assumption: comparing AWS, GCP, and Azure for a typical
web application workload at ~$500/month scale.

...
[Full decision support body]
...

AUDIT
ACTION: Select provider based on existing team expertise and
        primary region availability.
RISK: Cost estimates shift significantly with egress and
      reserved-instance commitments.
BASIS: Public pricing pages; cross-checked against known
       egress tiers.

​
Manual override

Start any message with exactly LOW, MID, or DEEP followed by whitespace to force that mode, bypassing automatic routing entirely. 
LOW what is the capital of France
MID explain the pros and cons of SSDs vs HDDs
DEEP compare these 3 cloud providers on cost and reliability
Manual override wins over structural detection. If you write DEEP but the message has no structural triggers, A.D.A.M. still enters DEEP — AUDIT is always ON in DEEP regardless of structural signals.
Boundary strictness. The override token must be followed by whitespace (space, tab, or newline) — not punctuation. MID o che bello is a valid override. MID, o che bello is not — the comma after the token breaks the boundary and the override is not recognised.

​
No-inertia rule

Mode is recomputed from scratch on every turn. A previous DEEP reply does not carry forward. The next message is routed independently based on its own structure and content. One-turn inheritance exception: A message with no whitespace that ends with ? inherits the previous exact LOW, MID, or DEEP for that one turn only. MODE: MID -> POSSIBLE DEEP is never inherited.
ScenarioBehaviour
Previous: DEEP. Next: normal messageNext is routed fresh — DEEP is not assumed
Previous: DEEP. Next: Really? (no whitespace, ends ?)Next inherits DEEP for one turn
Previous: MID -> POSSIBLE DEEP. Next: Really?No inheritance — POSSIBLE DEEP is never inherited

Build docs developers (and LLMs) love

Get started for freeTalk to us