DSPM Memory

Compress multi-turn LLM conversations by 80%+ while every constraint and decision survives.

100% critical retention across our benchmark suite — under impossibly small budgets, criticals are trimmed to their word floor and dropped only as a documented last resort.

---

The Problem

Long conversations eat your context window. Naive truncation drops the constraint from turn 3 that the entire system depends on. DSPM fixes this.

---

How It Works

DSPM converts each conversation turn into typed semantic patches — constraint, decision, code, entity, structure — and compresses them under a fixed token budget. Critical patches (constraints and decisions) are structurally protected: they survive compression even when everything else is trimmed.

Raw conversation (452 tokens, 18 turns)
        ↓
Semantic extraction → 28 patches, 18 critical
        ↓
7-stage compression pipeline
        ↓
Compressed context (249 tokens) — 100% of criticals intact

The 7-stage pipeline: dedup → slot fusion → delta encoding → causal pruning → utility scoring → shadow selection → adaptive budgeting.

---

Install

pip install dspm-memory

Requires Python 3.9+. Works with any OpenAI-compatible LLM provider.

Optional: semantic ranking

pip install "dspm-memory[semantic]"

Adds sentence-transformer-based query alignment for better non-critical patch ranking (downloads PyTorch and a small embedding model on first use). Without it, DSPM falls back to a neutral alignment score — criticals are fully protected either way.

---

Quickstart

from openai import OpenAI
from dspm import DSPMMemory

# Works with OpenAI, Groq, Together, Ollama, or any OpenAI-compatible endpoint
llm = OpenAI(
    api_key="sk-...",
    # base_url="https://api.groq.com/openai/v1"  # uncomment for Groq
)

# budget=60 so compression is visible even in this small example;
# use 200-300 for real conversations
memory = DSPMMemory(budget=60, llm_client=llm, model="gpt-4o-mini")

memory.add_turn("user", "Building a payment API. Hard rules: PCI-DSS compliant, max fee 0.5%, deadline Friday.")
memory.add_turn("assistant", "PCI-DSS needs tokenized card storage and quarterly ASV scans. Tracking the 0.5% fee cap.")
memory.add_turn("user", "Webhook timeout must be 30 seconds.")
memory.add_turn("assistant", "Webhook timeout set to 30 seconds with automatic retries.")
memory.add_turn("user", "Change the webhook timeout to 10 seconds instead — 30 is too slow.")
memory.add_turn("assistant", "Updated: webhook timeout is now 10 seconds; the 30s setting is superseded.")

context = memory.get_context(query="What are the hard requirements?")
print(context)
print(memory.stats)

Output:

[DEC] Track the 0.5% fee cap.
[CON] PCI-DSS requires tokenized card storage and ASV
[DEC] webhook timeout updated to 10 seconds
[CON] PCI-DSS compliant, max fee 0.5%, deadline Friday
{'turns': 6, 'total_patches': 4, 'critical_total': 4, 'critical_selected': 4, 'crr': 100, 'budget': 60, 'raw_tokens': 61, 'context_tokens': 56, 'trr': 9}

Note the revision: only the 10-second entry remains — the 30s setting is fully superseded. The fee cap and deadline survived compression, trr: 9 shows real token reduction even at this small scale, and crr: 100 throughout.

---

Persistence: Memory Across Chats and Sessions

Save the notebook when a chat ends, load it when the next one starts — Chat 2 remembers Chat 1, and revisions supersede old values across sessions:

# Chat 1 — Monday
memory = DSPMMemory(budget=250, llm_client=llm, model="gpt-4o-mini")
memory.add_turn("user", "Building a budgeting app. Hard rules: must work offline, deadline Oct 20.")
# ... chat ...
memory.save("user_dhruv.json")

# Chat 2 — Thursday, new process, fresh start
memory = DSPMMemory(budget=250, llm_client=llm, model="gpt-4o-mini")
memory.load("user_dhruv.json")
memory.get_context(query="What were my hard rules?")
# → [CON] must work offline
# → [CON] deadline Oct 20        ← recalled from Chat 1, zero API cost

# Revisions work across chats too:
memory.add_turn("user", "The deadline moved to November 5.")
# → Oct 20 is superseded. November 5 replaces it.

Save files are portable JSON, written atomically (a crash mid-save can't corrupt the notebook), and merge-on-load means saved revisions supersede stale values. One file per user = each person's long-term memory.

---

The Guarantee

[CON] and [DEC] patches are structurally protected:

• Never dropped by deduplication, fusion, or pruning
• Under budget pressure, payloads are trimmed numbers-first — and units bind to their numbers (30 seconds stays 30 seconds, never just 30)
• When a constraint is revised mid-conversation (e.g. TTL 60s → 300s), the new value supersedes the old — including cross-type revisions (a constraint revised by a decision) and terse patches
• A critical is only dropped as a last resort: every critical already at its 2-word floor and budget still cannot hold them
• Transient API failures (rate limits, 5xx) are retried automatically with backoff — three attempts before the error surfaces

Ablation result: Removing the shadow-selection mechanism collapses CRR from 100% to 37.9%, isolating the guarantee to a single identifiable component.

---

Results

Tested across 7 domains × 40 turns each:

CRR = Critical Retention Rate. TRR = Token Reduction Ratio.

Provenance: measured on 7 hand-authored 40-turn technical dialogues (API design, ML ops, IoT, security IR, supply chain, clinical workflow, project management); extraction and judging by GLM via OpenRouter (the since-renamed "ox-alpha" model). Numbers vary by extraction model and domain — check your own with memory.stats.

---

Supported Providers

# OpenAI
llm = OpenAI(api_key="sk-...")

# Groq (free tier available)
llm = OpenAI(base_url="https://api.groq.com/openai/v1", api_key="gsk-...")

# Together AI
llm = OpenAI(base_url="https://api.together.xyz/v1", api_key="...")

# Ollama (local, no key needed)
llm = OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")

---

API Reference

DSPMMemory(budget, llm_client, model)

The budget can be changed at any time:

memory.budget = 100        # takes effect on the next get_context() call
memory.set_budget(100)     # explicit alternative — same effect

Methods

Properties

---

Paper

DSPM: A Critical-Retention Approach to Long-Context Memory Compression for LLM Conversations
Dhruv Dubey, 2026
Zenodo: 10.5281/zenodo.19438636

---

Version History

---

License

MIT © 2026 Dhruv Dubey