# Chordian Docs

> Sovereign AI memory infrastructure for the enterprise. Chordian unifies enterprise knowledge into a typed, temporal memory that answers with citations — hosted by Chordian or running entirely on the customer's own infrastructure.

Every page below is also available as markdown by appending `.md` to its URL.

## Guidelines

- [Overview](https://docs.uni-flow.ai/guidelines.md): What Chordian is, who it is for, and how to find your way around these docs.
- [Best practices](https://docs.uni-flow.ai/guidelines/best-practices.md): What separates a memory that works from one that quietly degrades.
- [Knowledge graph](https://docs.uni-flow.ai/guidelines/concepts/knowledge-graph.md): Typed entities, typed relationships, and facts that know when they were true.
- [Knowledge quality](https://docs.uni-flow.ai/guidelines/concepts/knowledge-quality.md): Contradiction detection, health scoring and compiled syntheses.
- [Memory model](https://docs.uni-flow.ai/guidelines/concepts/memory-model.md): Tenants, workspaces, memories, departments and spaces — and which boundary does what.
- [Retrieval](https://docs.uni-flow.ai/guidelines/concepts/retrieval.md): How a question becomes a cited answer — four engines, reranking, and synthesis.
- [Spaces](https://docs.uni-flow.ai/guidelines/concepts/spaces.md): Scoped containers that give a project, team or agent its own slice of memory.
- [FAQ](https://docs.uni-flow.ai/guidelines/faq.md): Direct answers to the questions that come up in most evaluations.
- [Connectors](https://docs.uni-flow.ai/guidelines/integrations/connectors.md): Sync Notion, Drive, Gmail, Confluence, HubSpot and 350+ other sources.
- [MCP and agents](https://docs.uni-flow.ai/guidelines/integrations/mcp.md): Connect Claude Code, Cursor and other assistants directly to a memory or space.
- [Webhooks](https://docs.uni-flow.ai/guidelines/integrations/webhooks.md): React to ingestion, entity and quality events in your own systems.
- [Platform features](https://docs.uni-flow.ai/guidelines/platform/features.md): What ships today across ingestion, retrieval, quality, governance and access.
- [Memory types](https://docs.uni-flow.ai/guidelines/platform/memory-types.md): Four ways to run a Chordian memory, what each one buys you, and how to choose.
- [Credits and limits](https://docs.uni-flow.ai/guidelines/platform/token-economy.md): One currency for metered work, two independent gates, and how usage is measured.
- [Quickstart](https://docs.uni-flow.ai/guidelines/quickstart.md): Create a memory, add a document, and get a cited answer — in about five minutes.
- [Scenarios](https://docs.uni-flow.ai/guidelines/scenarios.md): What the platform adds up to for revenue, support, engineering and research teams.
- [Compliance](https://docs.uni-flow.ai/guidelines/security/compliance.md): What the architecture makes structural, what remains procedural, and what is missing.
- [Data perimeter](https://docs.uni-flow.ai/guidelines/security/data-perimeter.md): Secrets and PII detected, classified and masked before storage and before inference.
- [Security overview](https://docs.uni-flow.ai/guidelines/security/overview.md): The nine layers a request passes through, and what each one is defending.
- [Data sovereignty](https://docs.uni-flow.ai/guidelines/security/sovereignty.md): Run a memory on infrastructure Chordian does not own — and cannot read.
- [Tenant isolation](https://docs.uni-flow.ai/guidelines/security/tenant-isolation.md): How the boundary between organisations is enforced — and how it is proven.

## Platform API

- [Platform API](https://docs.uni-flow.ai/api.md): The control plane — workspaces, memories, credentials, spaces, usage and audit.
- [Authentication](https://docs.uni-flow.ai/api/authentication.md): Key types, scoping, rotation and revocation — and choosing the right credential.
- [Conventions](https://docs.uni-flow.ai/api/conventions.md): Identifiers, pagination, idempotency, timestamps and versioning.
- [Errors](https://docs.uni-flow.ai/api/errors.md): Status codes, the error shape, and what to do about each one.
- [Getting started](https://docs.uni-flow.ai/api/getting-started.md): Your first authenticated request against the platform control plane.
- [Rate limits and quotas](https://docs.uni-flow.ai/api/rate-limits.md): Per-credential limits, per-memory limits, and how to back off correctly.
- [Create an API key](https://docs.uni-flow.ai/api/reference/api-keys/createApiKey.md): Mints a key and registers it with the gateway.
- [List API keys](https://docs.uni-flow.ai/api/reference/api-keys/listApiKeys.md): Key metadata only.
- [Revoke an API key](https://docs.uni-flow.ai/api/reference/api-keys/revokeApiKey.md): Revokes the credential at the gateway **first**, then tombstones the record — so there is no window where a revoked key still authenticates.
- [Rotate an API key](https://docs.uni-flow.ai/api/reference/api-keys/rotateApiKey.md): Mints a replacement inheriting the original's name and scopes, then soft-revokes the old one.
- [List audit events](https://docs.uni-flow.ai/api/reference/audit/listAuditEvents.md): Immutable, append-only record of privileged and destructive actions.
- [Create a memory](https://docs.uni-flow.ai/api/reference/memories/createMemory.md): Creates a memory box.
- [Delete a memory](https://docs.uni-flow.ai/api/reference/memories/deleteMemory.md): Destroys the memory and its content.
- [Get a memory](https://docs.uni-flow.ai/api/reference/memories/getMemory.md): Get a memory
- [Get memory profile](https://docs.uni-flow.ai/api/reference/memories/getMemoryProfile.md): Composition of the memory: document and entity counts, departments, connected sources and last activity.
- [List memories](https://docs.uni-flow.ai/api/reference/memories/listMemories.md): Every memory in the tenant, across all workspaces.
- [Create a space](https://docs.uni-flow.ai/api/reference/spaces/createSpace.md): A space is a named, scoped container inside a memory — one per project, team or integration.
- [Delete a space](https://docs.uni-flow.ai/api/reference/spaces/deleteSpace.md): Delete a space
- [Get a space](https://docs.uni-flow.ai/api/reference/spaces/getSpace.md): Get a space
- [List spaces](https://docs.uni-flow.ai/api/reference/spaces/listSpaces.md): List spaces
- [Add a turn](https://docs.uni-flow.ai/api/reference/threads/addThreadTurn.md): Adds a question to the thread.
- [Create a thread](https://docs.uni-flow.ai/api/reference/threads/createThread.md): Create a thread
- [List threads](https://docs.uni-flow.ai/api/reference/threads/listThreads.md): List threads
- [Get credit balance](https://docs.uni-flow.ai/api/reference/usage/getCreditBalance.md): `available` is what you can actually spend — it is `balance` minus any credits currently reserved against in-flight work.
- [Get LLM token usage](https://docs.uni-flow.ai/api/reference/usage/getLlmUsage.md): Token and cost accounting for every LLM call attributable to your tenant.
- [Get memory usage](https://docs.uni-flow.ai/api/reference/usage/getMemoryUsage.md): Stored bytes and token consumption, per day and per department.
- [Register a webhook](https://docs.uni-flow.ai/api/reference/webhooks/createWebhook.md): Deliveries are signed with HMAC-SHA256 and retried on failure at 1s / 2s / 4s / 8s / 30s, then dead-lettered after five attempts.
- [List webhooks](https://docs.uni-flow.ai/api/reference/webhooks/listWebhooks.md): List webhooks
- [Set the active workspace](https://docs.uni-flow.ai/api/reference/workspaces/activateWorkspace.md): Switches the tenant's active workspace.
- [Create a workspace](https://docs.uni-flow.ai/api/reference/workspaces/createWorkspace.md): Creates a workspace and provisions its primary shared memory.
- [Delete a workspace](https://docs.uni-flow.ai/api/reference/workspaces/deleteWorkspace.md): Blocked while the workspace still owns memories, and the default workspace can never be deleted.
- [Get a workspace](https://docs.uni-flow.ai/api/reference/workspaces/getWorkspace.md): Get a workspace
- [List workspaces](https://docs.uni-flow.ai/api/reference/workspaces/listWorkspaces.md): Returns every workspace in the tenant, plus which one is currently active.
- [Update a workspace](https://docs.uni-flow.ai/api/reference/workspaces/updateWorkspace.md): Update a workspace
- [Webhooks](https://docs.uni-flow.ai/api/webhooks.md): Register endpoints, verify signatures and handle redelivery correctly.

## Memory API

- [Memory API](https://docs.uni-flow.ai/memory-api.md): Ingest, retrieve and reason over your organisation's knowledge.
- [Getting started](https://docs.uni-flow.ai/memory-api/getting-started.md): Ingest content and retrieve a cited answer from it.
- [Ingestion](https://docs.uni-flow.ai/memory-api/guides/ingestion.md): Routing, files, deduplication and what happens after content arrives.
- [Knowledge quality](https://docs.uni-flow.ai/memory-api/guides/knowledge-quality.md): Working with contradictions, health scores, syntheses and merge candidates.
- [Retrieval](https://docs.uni-flow.ai/memory-api/guides/retrieval.md): Filters, scoping, streaming and handling citations well.
- [Memory model](https://docs.uni-flow.ai/memory-api/memory-model.md): Scoping a request, choosing a host, and which credential to use.
- [Delete an upload](https://docs.uni-flow.ai/memory-api/reference/ingestion/deleteUpload.md): Removes the file, its chunks and its vectors.
- [Get an upload](https://docs.uni-flow.ai/memory-api/reference/ingestion/getUpload.md): Get an upload
- [Ingest content](https://docs.uni-flow.ai/memory-api/reference/ingestion/ingestContent.md): The universal text ingestion endpoint.
- [List uploaded files](https://docs.uni-flow.ai/memory-api/reference/ingestion/listUploads.md): List uploaded files
- [Re-index an upload](https://docs.uni-flow.ai/memory-api/reference/ingestion/reindexUpload.md): Re-runs extraction and indexing.
- [Upload a file](https://docs.uni-flow.ai/memory-api/reference/ingestion/uploadFile.md): Accepts PDF, DOCX, XLSX, CSV, TXT, Markdown and images.
- [Traverse from an entity](https://docs.uni-flow.ai/memory-api/reference/knowledge-graph/exploreGraph.md): Walks outward from an entity.
- [Get an entity](https://docs.uni-flow.ai/memory-api/reference/knowledge-graph/getEntity.md): Get an entity
- [Get an entity timeline](https://docs.uni-flow.ai/memory-api/reference/knowledge-graph/getEntityTimeline.md): Every recorded value of every attribute, with both timestamps that matter: when the fact was **true** (`valid_at`) and when Chordian **learned** it (`recorded_…
- [Get the knowledge graph](https://docs.uni-flow.ai/memory-api/reference/knowledge-graph/getGraphView.md): Nodes and edges for the memory, ready to render.
- [Get memory health](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/getMemoryHealth.md): A report card for the **quality of the knowledge** — not its security.
- [Get an entity synthesis](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/getSynthesis.md): A compiled, always-current profile for one entity — overview, current relationship, key contacts, recent activity, risks and likely outcome.
- [List contradictions](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/listContradictions.md): Facts your sources disagree about.
- [List merge candidates](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/listMergeCandidates.md): Entities that look like duplicates but scored below the auto-merge threshold — "Jana Müller" and "Jana Mueller", for instance.
- [Merge two entities](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/mergeCandidate.md): Merges the pair non-destructively — the losing entity is linked with a `SAME_AS` edge rather than deleted, so provenance survives.
- [Recompile a synthesis](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/refreshSynthesis.md): Recompile a synthesis
- [Resolve a contradiction](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/resolveContradiction.md): Writes your decision back to the graph and records who made it.
- [Run health scoring now](https://docs.uni-flow.ai/memory-api/reference/knowledge-quality/runHealthLint.md): Queues an immediate re-score instead of waiting for the nightly job.
- [Teach the memory](https://docs.uni-flow.ai/memory-api/reference/learning/learn.md): Feed corrections back so the memory improves instead of repeating a mistake.
- [List indexed sources](https://docs.uni-flow.ai/memory-api/reference/search/listMemorySources.md): Every source that has actually contributed indexed content, with document counts.
- [Search memory](https://docs.uni-flow.ai/memory-api/reference/search/memorySearch.md): Runs the full retrieval pipeline and returns a synthesised answer with citations.
- [Stream a search answer](https://docs.uni-flow.ai/memory-api/reference/search/memorySearchStream.md): Same pipeline as `POST /search`, delivered as Server-Sent Events.
- [Capture a conversation](https://docs.uni-flow.ai/memory-api/reference/spaces/captureConversation.md): Stores an LLM conversation in the space so the organisation keeps what was asked and answered — including in tools Chordian does not host.
- [List conversations across spaces](https://docs.uni-flow.ai/memory-api/reference/spaces/listAllConversations.md): Every captured conversation in the memory, filterable by originating tool — one place to see what the organisation has been asking, whichever assistant it was…
- [Ask a space](https://docs.uni-flow.ai/memory-api/reference/spaces/spaceAsk.md): Runs an agent scoped to one space.

## Optional

- [OpenAPI — Platform API](https://docs.uni-flow.ai/openapi/platform.yaml)
- [OpenAPI — Memory API](https://docs.uni-flow.ai/openapi/memory.yaml)