context-compiler 0.4.4

Deterministic conversational state engine for LLM applications.

Context Compiler

A deterministic directive engine that converts explicit user instructions into structured conversational state for LLM applications.

Modern language models reason well but are unreliable at maintaining consistent state across interactions.

Corrections compete with earlier statements, constraints disappear, and long conversations accumulate contradictions.

The Context Compiler introduces a deterministic state layer that governs authoritative conversational state independently of the model.

The model performs reasoning and generation while the compiler manages facts and constraints. Once accepted, directives remain authoritative until explicitly corrected or reset.

Why “Compiler”?

Context Compiler treats explicit user directives as inputs to a deterministic process.

Instead of relying on the LLM to remember constraints across a conversation, user instructions are compiled into structured state before the model runs.

The idea is similar to a traditional compiler: user directives are translated into a structured representation that the rest of the system can rely on.

Installation

• Python 3.11+
• pip install context-compiler
• Dev/test: uv sync --group dev and uv run pytest
• Examples: see examples/README.md
• Demonstrations: see demos/README.md

---

10-Second Example

User sets a constraint once:

User: don't use peanuts

Outcome: prohibited items now include "peanuts".

Later in the conversation:

User: how should I make this curry?

The host supplies the authoritative state to the model so the constraint persists across turns.

---

Architecture

User Input
     │
     ▼
Context Compiler
     │
     ▼
Decision
     │
     ▼
Host Application
 ├─ clarify → ask user
 ├─ passthrough → call LLM
 └─ update → call LLM with compiled state

The compiler governs authoritative conversational state and never calls the LLM. The host decides whether to call the model based on the returned Decision.

---

Decision API

Each user message produces a Decision.

class Decision(TypedDict):
    kind: Literal["passthrough", "update", "clarify"]
    state: dict | None
    prompt_to_user: str | None

Meaning:

kind	host behavior
passthrough	forward user input to LLM
update	forward input with updated state
clarify	show prompt_to_user and do not call the LLM

Host Integration Example

engine = create_engine()
decision = engine.step(user_input)

if decision["kind"] == "clarify":
    show_to_user(decision["prompt_to_user"])
else:
    state = decision["state"] or engine.state
    messages = build_messages(state, user_input)
    render(call_llm(messages))

API Reference

API	Description
create_engine(state=None)	Create a new compiler engine; optional state provides initial authoritative state (validated/canonicalized).
step(user_input)	Parse one user turn and return a deterministic Decision.
compile_transcript(messages)	Replay a transcript from a fresh engine and return either final state or a confirmation prompt.
engine.apply_transcript(messages)	Replay a transcript onto the current engine state and return either final state or a confirmation prompt.
engine.state	Read current authoritative in-memory state snapshot.
get_focus_value(state)	Read the current focus value from a state snapshot.
get_prohibited_items(state)	Read prohibited items from a state snapshot.
export_json()	Export current state as JSON for persistence/transport.
import_json(payload)	Load/restore state from exported JSON payload.

---

State Model

The compiler maintains an authoritative state snapshot. Hosts should treat this state as structured application data and avoid coupling to internal field names or nested layout.

State Access and Persistence

Hosts can provide initial state at engine creation (create_engine(state=...)), read current in-memory state via engine.state, and persist/restore via export_json() and import_json(). Semantic state mutations occur through directives processed by step(). Storage is managed by the host application.

Use the returned state snapshot as structured host input for prompt construction, policy enforcement, or replay/storage workflows. For host code that needs typed reads without direct nested key lookups, use get_focus_value(state) and get_prohibited_items(state).

Transcript Replay

Transcript replay compiles conversational history by reusing the same deterministic directive path:

• Only messages with role == "user" are processed.
• Assistant/system/non-user messages are ignored.
• Replay calls step() for each user message in order.
• Replay stops on the first clarification and returns a confirmation prompt.
• compile_transcript(messages) starts from a fresh engine.
• engine.apply_transcript(messages) applies replay onto the current engine state.

Fact Schema

The current behavior includes one exclusive focus value. This demonstrates deterministic fact replacement and correction behavior. Richer schemas may be introduced in future releases.

State Properties

• Facts are exclusive (last write wins)
• Policies are additive
• No inference or semantic reasoning

Identical input sequences always produce identical compiler state. LLM responses may still vary unless deterministic decoding is used by the host.

Example:

User: use tofu
User: use corn oil

Result: the current focus value becomes "corn oil"

Because the focus value is exclusive (last write wins), later use ... directives replace earlier values.

This may differ from human expectations, where the intent may be interpreted as additive (e.g., ingredient + cooking medium). The current schema models a single focus value. See issue #45 for discussion.

---

Directive Examples

Hard negative directive:

User: don't use peanuts

Result: prohibited items include "peanuts".

Fact configuration:

User: use vegetarian curry

State update: the current focus value becomes "vegetarian curry"

Correction:

User: actually vegan curry

Result: the current focus value becomes "vegan curry"

Ambiguous mutation:

User: no use peanuts

Compiler response:

Decision.kind = "clarify"

No state mutation occurs until confirmation.

Reset Commands

Two explicit reset commands are supported:

• reset policies clears prohibited items but preserves the current focus value
• clear state resets the full state to initial values

Example:

• If current focus is "vegetarian curry" and prohibited items include "peanuts":
• after reset policies, prohibited items are empty and focus remains "vegetarian curry".
• after clear state, both focus and prohibited items return to initial defaults.

---

Examples

Integration examples are available in the examples/ directory.

See examples/README.md for walkthroughs.

---

Quickstart

Run the interactive REPL:

context-compiler

Run an example:

python examples/01_persistent_guardrails.py

Run tests:

uv run pytest

---

Guarantees

• State changes only through explicit user directives or confirmation.
• Identical input sequences produce identical compiler state.
• Model responses never modify compiler state.
• Ambiguous directives trigger clarification instead of changing state.

These invariants are verified through behavioral tests and Hypothesis-based property tests.

---

Design Notes

More detailed design and milestone documents are available in:

• Project overview
• Directive grammar specification

---

License

Apache-2.0.