NocturnusAI

The context engineering engine for AI agents: send only what changed.

Before / After

# ❌ Without NocturnusAI — replay everything, every turn
messages = system_prompt + full_history + tool_outputs   # ~1,259 tokens/turn
response = llm(messages)                                 # $13,600/mo at scale

# ✅ With NocturnusAI — send only what changed
ctx = nocturnus.process_turns(raw_turns)                 # extract → infer → delta
messages = system_prompt + ctx.briefing_delta             # ~221 tokens/turn
response = llm(messages)                                 # $2,400/mo. Same accuracy.

The Numbers

Measured on live APIs. 15-turn product support conversation. Real usage.input_tokens counts. Run it yourself.

Claude Opus 4: 5.7× reduction. Gemini 2.0 Flash: 10.0×. Full calculations.

Install

pip install nocturnusai          # Python
npm install nocturnusai-sdk      # TypeScript
docker run -p 9300:9300 ghcr.io/auctalis/nocturnusai:latest  # Docker

Or use the setup wizard:

curl -fsSL https://raw.githubusercontent.com/Auctalis/nocturnusai/main/install.sh | bash

Why Developers Star This Repo

• Reproducible token reduction — benchmark in the repo, methodology published, run it against your own workload
• Deterministic inference — same query, same result, every time. No embedding drift, no cosine similarity lottery
• Truth maintenance — retract a fact, all derived conclusions auto-retract. No stale context, no hallucination on operational state
• Plugs into existing stacks — LangChain, LlamaIndex, CrewAI, AutoGen, MCP, Vercel AI SDK, OpenAI Agents SDK, Mastra
• Benchmarkable against naive replay — numbers derived, not invented. Every claim traces to a notebook cell

Framework Quickstarts

How It Works

Three steps. Every turn.

1. Extract — raw conversation turns → structured facts via LLM extraction
2. Infer — backward-chaining logical inference finds only the facts reachable from the agent's current goal
3. Return the delta — a briefingDelta containing only what changed since the last turn

This is not vector search. It is not summarization. It is deterministic inference on a logic engine — Hexastore indexing, backward chaining, and truth maintenance.

The Working Loop

LLM required for natural-language turns. The examples below send raw text turns through an LLM to extract structured facts. If you start the server without an LLM provider, natural-language turns will return zero facts. See Quick Start for setup options, or use predicate syntax (e.g., "customer_tier(acme_corp, enterprise)") which works without any LLM.

1. First reduction: POST /context

curl -X POST http://localhost:9300/context \
  -H 'Content-Type: application/json' \
  -H 'X-Tenant-ID: default' \
  -d '{
    "turns": [
      "user: Customer says they are enterprise and blocked on SLA credits.",
      "tool: CRM says account is Acme Corp with a 2M ARR contract.",
      "agent: Last week support promised to review SLA eligibility.",
      "tool: Billing note says renewal is due next month."
    ],
    "maxFacts": 12
  }'

2. Goal-driven pass: POST /memory/context

curl -X POST http://localhost:9300/memory/context \
  -H 'Content-Type: application/json' \
  -H 'X-Tenant-ID: default' \
  -d '{
    "goals": [{"predicate":"eligible_for_sla","args":["acme_corp"]}],
    "maxFacts": 12,
    "sessionId": "ticket-42"
  }'

3. Later turns: POST /context/diff

curl -X POST http://localhost:9300/context/diff \
  -H 'Content-Type: application/json' \
  -H 'X-Tenant-ID: default' \
  -d '{"sessionId": "ticket-42", "maxFacts": 12}'

Returns only added and removed entries between snapshots.

4. End of thread: POST /context/session/clear

curl -X POST http://localhost:9300/context/session/clear \
  -H 'Content-Type: application/json' \
  -H 'X-Tenant-ID: default' \
  -d '{"sessionId":"ticket-42"}'

Choose Your Surface

from nocturnusai import SyncNocturnusAIClient

with SyncNocturnusAIClient("http://localhost:9300") as client:
    ctx = client.process_turns(
        turns=[
            "user: Customer says they are enterprise and blocked on SLA credits.",
            "tool: CRM says account is Acme Corp with a 2M ARR contract.",
        ],
        scope="ticket-42",
        session_id="ticket-42",
    )

    diff = client.diff_context(session_id="ticket-42", max_facts=12)
    client.clear_context_session("ticket-42")

    print(ctx.briefing_delta)

import { NocturnusAIClient } from 'nocturnusai-sdk';

const client = new NocturnusAIClient({
  baseUrl: 'http://localhost:9300',
  tenantId: 'default',
});

const ctx = await client.processTurns({
  turns: [
    'user: Customer says they are enterprise and blocked on SLA credits.',
    'tool: CRM says account is Acme Corp with a 2M ARR contract.',
  ],
  scope: 'ticket-42',
  sessionId: 'ticket-42',
});

const diff = await client.diffContext({ sessionId: 'ticket-42', maxFacts: 12 });
await client.clearContextSession('ticket-42');
console.log(ctx.briefingDelta);

{
  "mcpServers": {
    "nocturnus": {
      "url": "http://localhost:9300/mcp/sse",
      "transport": "sse"
    }
  }
}

What Lives Behind The Workflow

When you do need backend mechanics, NocturnusAI provides them:

• Deterministic fact and rule storage
• Backward-chaining inference with proof chains
• Truth maintenance and contradiction handling
• Temporal facts with ttl, validFrom, and validUntil
• Multi-tenancy via X-Database and X-Tenant-ID
• MCP, REST, Python SDK, TypeScript SDK, and CLI surfaces over the same engine

Quick Start

Docker (fastest)

docker run -d --name nocturnusai -p 9300:9300 \
  --restart unless-stopped \
  -v nocturnusai-data:/data \
  ghcr.io/auctalis/nocturnusai:latest

curl http://localhost:9300/health   # Verify it's running

Docker with Ollama (enables natural-language extraction)

docker run -d --name nocturnusai -p 9300:9300 \
  --add-host=host.docker.internal:host-gateway \
  -e LLM_PROVIDER=ollama \
  -e LLM_MODEL=granite3.3:8b \
  -e LLM_BASE_URL=http://host.docker.internal:11434/v1 \
  -e EXTRACTION_ENABLED=true \
  ghcr.io/auctalis/nocturnusai:latest

From this repo

make up-ollama && make smoke

CLI

nocturnusai                                # Interactive REPL
nocturnusai -e "context 10"               # Salience-ranked working set
nocturnusai -e "compress"                 # POST /memory/compress
nocturnusai -e "cleanup 0.05"             # POST /memory/cleanup

Documentation

Full docs: nocturnus.ai

Docker Compose (advanced)

git clone https://github.com/Auctalis/nocturnusai.git && cd nocturnusai

make up                                        # Server using .env.example defaults
make up-ollama                                 # + Ollama (reuses host or starts bundled)
make up-monitoring                             # + Prometheus + Grafana
make smoke                                     # Verify health + context endpoint

Build from Source

Requires JDK 17+.

./gradlew :nocturnusai-server:run              # HTTP server on :9300
./gradlew :nocturnusai-cli:run                 # Interactive REPL (JVM)
./gradlew :nocturnusai-cli:nativeCompile       # Build native binary
./gradlew test                                 # Full test suite

Contributing

See CONTRIBUTING.md. Issues labelled good first issue are good entry points.

Security

Report vulnerabilities privately via GitHub Security Advisories. See SECURITY.md.

License

Business Source License 1.1 (SPDX: BUSL-1.1). Free for internal use — including internal production — inside your own organization. Offering NocturnusAI or substantial functionality as a product/hosted service to third parties requires a commercial license (licensing@nocturnus.ai). Converts to Apache 2.0 on 2030-02-19. See LICENSE and DISCLAIMER.md.

LEGAL & SAFETY NOTICE

NocturnusAI is a deterministic reasoning engine, but its output is only as reliable as the facts provided to it.

1. No Warranty of Truth. "Verified" refers to logical consistency of inference, not accuracy of real-world claims.
2. Not for Autonomous High-Stakes Decisions. Do not use this engine for unsupervised medical, financial, legal, or physical-safety decisions without an independent human verification step.
3. Logic Layer Only. NocturnusAI provides information and inference; it does not execute actions.
4. No Liability. See DISCLAIMER.md and LICENSE.