Veracium

Veracium is a provenance-aware memory plug-in for agentic systems — durable, per-user memory that resists the injection and confabulation failures that plague naive agent memory. Provenance means every fact tracks who said it: a claim from an email your agent merely read can never become a "fact" it asserts. It remembers what the user said, past interactions, and what worked — and it remembers where each of those came from.

Veracium is the production distillation of an evaluation-driven research project (agent-memory): every design choice below traces to a measured finding, and the research's synthetic-corpus harness is reused as the regression suite.

Research: the evaluation instrument behind those findings — a longitudinal benchmark for agent memory — is described in Q. Spencer, "Ground Truth First: A Longitudinal Evaluation Instrument for Agent Memory, and the Tenure Crossover in Memory-Architecture Rankings" (arXiv:2607.21962, 2026).

Why it's shaped this way

• Typed graph + dated episodes are the store of record. Entity facts live as relational edges (with unforgeable provenance); interaction history lives as dated episodes. A curated "wiki" view is compiled from them and cached — never the source of truth. (The layered design won on both short and 9-week horizons; flat stores each failed one regime.)
• Supersession, never erasure. Functional facts (preference, employer, deadline) keep one current value with the prior value retained as history — "what did X used to be?" stays answerable. (The category commercial memory systems handle worst; Veracium's strongest.)
• Representation is a security control. Third-party claims (received email, external docs) are quarantined structurally — stored as third_party_claim edges with the claimant as subject, never as user facts. Content-type quarantine catches obligation/debt/renewal claims regardless of how plausible they look. (Held against a full plausibility ladder incl. contact-impersonation.) What this is, and is not: the store governs retention, retrieval, description and recommendation — what it will say and how it labels it. It does not instantiate, authorize or execute anything; those belong to the host's harness. Without an exclusive harness path that consumes the labels, Veracium's trust classes are advisory labelling, not enforcement.
• Bring your own model. Veracium never owns your API keys or model choice; it calls a Complete callable you supply. A reference Anthropic provider ships in the box.
• Embedded by default. Zero external services: one SQLite file. Swap in Neo4j/Postgres later via the Store interface.

Install

pip install "veracium[anthropic]"   # core + the reference LLM provider

Extras: [mcp] adds the MCP server, [dev] adds pytest. The core alone depends only on pydantic. To work from source instead:

git clone https://github.com/veracium-ai/Veracium.git && cd Veracium
pip install -e ".[anthropic,dev]"

Links: docs · veracium.ai · PyPI

Use (library)

from veracium import Memory, EvidenceAuthor, EvidenceContext
from veracium.llm.anthropic import AnthropicComplete

mem = Memory(llm=AnthropicComplete())   # or pass your own Complete callable

# Remember interactions. `author` says WHO wrote the event; `context` is
# your positive attestation of HOW you captured it. Without a context the
# content class floors to derived(THIRD_PARTY) — never assertable — so a
# host that means "I captured this first-hand" says so:
mem.remember("alice", "USER: I'm vegetarian and have a dog named Ollie.",
             context=EvidenceContext.direct())
mem.remember("alice", "From billing@scam: you owe $900.",
             author=EvidenceAuthor.THIRD_PARTY, event_type="email",
             source_id="billing-mailbox",   # third-party content needs a source id
             context=EvidenceContext.direct())

# Recall grounded, provenance-flagged context for a prompt.
ctx = mem.recall("alice", "suggest a lunch spot")
print(ctx.context)   # states the vegetarian constraint; the $900 "claim" is
                     # rendered under a never-assert flag, not as a fact.

No Anthropic API key? AnthropicComplete is just a convenience — Veracium calls any Complete callable you supply. To run without SDK/key setup, wrap a client you already have; examples/claude_cli_provider.py wraps the claude CLI as a drop-in provider (from claude_cli_provider import ClaudeCLIComplete), and examples/openai_provider.py wraps any OpenAI-compatible chat-completions API (OpenAI itself, vLLM, Ollama's /v1 endpoint) via OpenAIComplete — point it at a local server with OpenAIComplete(base_url=...) and override models with whatever model name your server serves.

Use (MCP)

veracium-mcp exposes remember / recall / answer / maintain tools to any MCP-compatible agent (Claude Desktop/Code, others) with no host-side Python. See docs/mcp.md for the config JSON and tool reference.

Documentation

Hosted docs: veracium-ai.github.io/Veracium

• examples/demo.ipynb — the scam-email injection demo, runnable end to end (open in Colab).
• examples/langchain_memory.py — Veracium as the long-term memory layer of a LangChain chat app (session-keyed hybrid: LangChain buffers recent turns, Veracium holds durable facts with provenance and quarantine; your existing LangChain model powers both sides).
• docs/concepts.md — the mental model: edges vs episodes vs the compiled wiki, provenance & authorship, quarantine, the abstention gate, lifecycle.
• docs/recipes.md — short copy-paste examples, one per capability (quarantine, mixed provenance, budgeted recall, portability, feedback verbs, audit, local models).
• docs/api.md — the public API: Memory, MemoryConfig, EvidenceAuthor, providing your own LLM callable or store.
• docs/mcp.md — running and registering the MCP server.
• docs/design-rationale.md — why there's no update()/delete(), no LLM-free extraction, no TTL purging — and what's genuinely on the roadmap.
• docs/telemetry.md — the opt-in, anonymous, content-free usage statistics (off by default).
• docs/diagnostics.md — opt-in error reporting: local-first error log, consented + redacted send.
• ROADMAP.md · CHANGELOG.md

Status

The validated layered design is implemented, tested (44 offline tests, plus opt-in live tiers: the acceptance eval and a real-corpus robustness harness), and passes its own research-claim bar (5/5, 0 injection asserts). Roadmap v0.1–v0.7 complete, plus opt-in telemetry, a self-check, consented error reporting, and an operation audit log. See ROADMAP.md.

License

MIT