Project memory and change control for AI-native delivery teams

Keep your team, your clients, and your agents on the same project truth.

Accord Book captures decisions and constraints from normal work — Slack, GitHub, uploads, voice, docs — detects conflicts before drift turns into rework, drafts client-safe updates, and proposes Git-backed context under human approval.

Talk to the AI guide — voice & chat

Ask about fit, self-hosting, MCP guardrails, or the founding pilot — live on accord-book.app.

Apply for pilot — 3 spots left Read the engineering blog

Built for boutique dev agencies — especially where the owner is technical lead, PM, and client translator. Sweet spot is often smaller teams (up to ~20 engineers); the platform scales beyond that.

Featured in Founderr Pulse — Jul 2026

Built and benchmarked

Three capabilities that are shipped, measured, and ready for pilot teams to put to work.

Conflict detection

100% precision, 92.86% recall across 86 scenarios on the May 2026 benchmark run — zero observed false positives. Surfaces review-worthy contradictions with provenance, keeps owners in control.

See the benchmark →

Shipped pilot capabilities

Owner arbitration, ingestion backlog visibility, per-install Slack outbound, agent preflight via MCP, and Git-backed spec generation are all live in the current release.

Read the architecture summary →

Production retrieval

Vector retrieval, lexical recovery, and supersedes-aware reranking reconstruct current project state — not just semantically similar text. Validated at 990ms p95 and 90% judged accuracy.

Read the benchmark →

A shared memory layer for delivery work that keeps changing

Most project drift starts the same way: the latest request lives in one place, the decision it conflicts with lives somewhere else, and the team only connects them after rework begins.

Ingest project evidence

Pull project inputs from Slack, GitHub, uploaded documents, URLs, and docs into a project-scoped memory with timestamps and provenance. No manual tagging required.

Reconstruct current state

Retrieval designed for changing project truth: vector search, lexical recovery, and supersedes-aware reranking find what is current — not just what sounds similar.

Surface review-worthy conflicts

Turn structured claims into conflict candidates and durable findings, then present the useful subset for owner review with supporting provenance.

Propose governed updates

Generate digests and Git-backed .q_context documentation updates as proposals. Humans review and publish what becomes authoritative — AI proposes, the team decides.

What changes when your team runs Accord Book

Accord Book replaces scattered, reactive coordination with a single governed project memory layer.

Without Accord Book	With Accord Book
Project truth scattered across tools	Project memory assembled into one scoped system
Stale decisions resurface as if current	Superseded context down-ranked and traceable
Client updates require manual synthesis	Digests draft from project memory — owners approve before sending
Spec drift accumulates silently in docs repos	`.q_context` updates proposed by PR — humans merge what becomes authoritative
Conflict review starts after rework begins	Review-worthy contradictions surface earlier, with provenance attached

From raw inputs to governed project truth

Accord Book is built for working loops where agents, humans, and clients all need to operate from the same project record.

✓

Owner conflict arbitration

Resolve, defer, and reopen — decisions written to project record

Shipped

✓

Daily digest + Slack outbound

Per-install Slack push, digest generation from project memory

Shipped

✓

Ingestion backlog visibility

Project-scoped view of what has been ingested and what is pending

Shipped

✓

Agent preflight via MCP

MCP tools and example configs for coding and voice agents

Shipped

✓

arc42 + ADR spec generation

Architecture docs proposed into .q_context/ via PR — merge to publish

Shipped

✓

Slack ingestion

OAuth, channel mapping, and message ingestion

Validated

✓

GitHub ingestion

App config, repo mapping, and push event ingestion

Validated

✓

Memory extraction with provenance

Project-scoped memories with full source traceability

Validated

✓

Access control + project membership

Invite acceptance creates correct project membership and roles

Validated

Agent preflight

Before an agent acts, it queries Accord Book for constraints, failed approaches, and relevant project history — via MCP, so any coding or voice agent can use it.

Owner arbitration

When new requests conflict with prior decisions or constraints, owners get a provenance-backed review surface with resolve, defer, and reopen actions. Decisions become part of the project record.

Client-safe digests

Accord Book drafts plain-language project summaries from actual memory and approved evidence — not from a generic chat recap. Owners review before anything reaches the client.

Use cases

Problem-first pages for agency owners evaluating team memory tools.

Team memory for AI-assisted client work

Shared record for humans, clients, and coding agents — not another chatbot session store.

Client context memory for agencies

Project-scoped decisions and provenance when you juggle multiple client engagements.

Data Sovereignty & Compliance

Built for agencies handling sensitive and European client IP

When you build for European enterprises, fintechs, or regulated sectors, standard multi-tenant AI tools fail strict confidentiality frameworks (such as GDPR Art. 28 and Germany's §203 StGB professional secrecy). Accord Book keeps your client's memory inside your own boundary.

Single-tenant isolation

One deployment per agency on your own Linux host via Docker Compose. No shared multi-tenant database where cross-tenant queries or leaky permissions can expose another client's architecture.

Zero telemetry & zero call-home

We operate no central data lake. We cannot read your specs, PRDs, Slack discussions, or Git commit history. What happens on your deployment stays on your host.

Direct BYOK provider custody

You supply your own LLM and embedding API keys (e.g., Azure OpenAI EU, zero-retention OpenAI, or Anthropic). Prompts route straight from your server to your model provider — Vector Intelligence is never in the data chain of custody.

Founding pilot cohort — June 2026

3 spots remaining

We are working directly with a small cohort of AI-native dev agencies to validate Accord Book in real delivery environments. Pilot teams get the full product, hands-on setup with the founders, and founding-cohort pricing that locks in before public launch.

Hands-on setup

We configure Accord Book for your project structure, input sources, and delivery workflow together. You are not reading docs alone.

Direct product access

Pilot teams work directly with the founders. Your real-project feedback shapes the roadmap before the public release.

Founding-cohort pricing

Pilot pricing locks in before the public launch rate. Teams keep their rate as Accord Book grows and the product expands.

Right-sized for your team

Designed for dev agencies doing AI-assisted delivery — sweet spot often up to ~20 engineers, with room to grow. BYOK for LLM and embeddings — no token margin passed to you.

See what the pilot includes →

Or email desk@vector-intelligence.io directly.