TGMS — Agent-Native Bi-Temporal Graph Management System

A temporal graph database whose query surface is built for LLM agents — and whose answers can be audited claim by claim.

Project page & blog: https://zxf-work.github.io/tgms/ · Paper: paper/main.pdf

LLM agents are unreliable at exactly the things temporal graph analytics requires: arithmetic, identifiers, and asserting only what the evidence shows. TGMS's answer is architectural — give the model no opportunity to do any of them:

• a bi-temporal property graph (valid time × transaction time) that distinguishes evolution ("the edge ended") from correction ("we were wrong"), so agents can answer "what did we believe on March 1?" — a question no snapshot or RAG system can express;
• 13 verified temporal operators (reachability over time-respecting paths, δ-motifs, snapshot diffs, burst detection, interval joins, …) — typed, deterministic, bounded, cost-guarded, exposed as tools (MCP or in-process); identifiers must come from a resolver, arithmetic from a compute operator;
• a Planner–Executor–Verifier loop: the LLM only plans and reports; plans are statically validated (including a grounding rule that makes fabricated identifiers impossible and output-field contracts that reject invented result paths), executed deterministically with content-addressed traces, and every claim in the written answer is machine-checked against the trace that produced it — including truncation taint, so "correct arithmetic over incomplete evidence" is caught too.

Does it work?

Dev-split campaign (CollegeMsg, open-source models served locally on one 24 GB GPU; details + receipts in docs/TECHNICAL_REPORT.md):

pooled EM, Qwen2.5-14B	TGMS	vector-RAG	static-graph RAG	text-to-Cypher
all task families	0.41	0.09	0.05	0.18
correction probes ("as of tt…")	0.67	0.00	0.00	0.00

• vs static-graph RAG: +36 points, paired-bootstrap 95% CI [0.18, 0.59]
• verifier fault injection: 500/500 injected false claims caught, 0 false positives; emitted answers carry an unsupported-claim rate of 0.000
• operators meet all latency targets at 1M events (snapshot 98 ms, diff 163 ms, reachability 63–244 ms)

Quickstart

# macOS note: if this repo sits in an iCloud-synced folder, keep the venv
# outside it (iCloud sets the hidden flag on .pth files and Python 3.12+
# silently skips them):  export UV_PROJECT_ENVIRONMENT=$HOME/.venvs/tgms
uv sync --extra agent
make test                     # 81 tests: property, oracle, metamorphic, e2e

# build a real store + task suite (downloads CollegeMsg from SNAP)
make data-collegemsg suite-collegemsg

# call one verified operator — no LLM needed
uv run tgms call temporal_reachability \
  '{"src": "n9", "window": {"t_a": 1082040961000000, "t_b": 1088000000000000}}' \
  --store stores/collegemsg

# verifier acceptance experiment (deterministic, no LLM)
uv run tgms eval c2 --store stores/collegemsg \
  --suite stores/suite-collegemsg/suite.json --mutants 500

With any OpenAI-compatible LLM endpoint (e.g. vllm serve Qwen/Qwen2.5-7B-Instruct):

uv run tgms ask "How many nodes can n9 reach between ... and ...?" \
  --store stores/collegemsg --model openai/Qwen/Qwen2.5-7B-Instruct \
  --api-base http://localhost:8000/v1 --html trace.html   # auditable trace page

bash scripts/run_webapp.sh    # interactive guided demo at localhost:8080

Interfaces

Surface	Entry point	What it's for
Python library	tgms.open(...), Agent(store, model=…).ask(…)	research code, notebooks
MCP server	tgms serve --store PATH	hand the verified toolbox to any MCP-capable agent
CLI	tgms ingest/synth/tasks/call/ask/bench/memory/eval	reproducibility
Trace viewer	tgms ask … --html trace.html	ask → answer → audit the evidence (static, self-contained HTML)
Demo GUI	tgms webapp … / scripts/run_webapp.sh	guided tour: operators → agent → tamper demo → time travel

Correctness

Every operator is verified against an independent brute-force oracle (500 randomized cases per operator; 96% line coverage in tgms/temporal/), plus metamorphic properties — diff composition and bi-temporal immutability: any result pinned to a past belief state is byte-identical before and after later corrections. The write path is property-tested over random assert/retract/correct interleavings, and the append-only event log replays into either backend with identical store digests. Process rules (test ownership, decision log, determinism receipts) are enforced in CI — see CONTRIBUTING.md and docs/DECISIONS.md.

Layout

tgms/core       clock, bi-temporal data model, error taxonomy
tgms/storage    StorageAdapter ABC, Kùzu + DuckDB backends, event log, TCSR index
tgms/temporal   operator algebra O1–O13 + brute-force oracle
tgms/tools      tool schemas, MCP server / ToolRouter, trace viewer, demo GUI
tgms/agent      plan IR, planner, executor, verifier, reporter, memory
tgms/data       dataset loaders (SHA-256 pinned) + synthetic generator
tgms/eval       task suites, baselines, matrix harness, metrics, fault injection

Datasets are never bundled: loaders download from source (SNAP) and pin SHA-256 manifests. See docs/TECHNICAL_REPORT.md for design, positioning, measurements, and roadmap.

License

Apache-2.0 — see LICENSE. Cite via CITATION.cff.