Skip to main content

Turn Any GitHub Repo Into Interactive Docs & DB Diagrams

Paste your repository link to instantly generate API & page docs, visualize database relations, and chat with your codebase.

Try a live example:

GitHub integration  ·  No credit card required

connected
Schema Visualizersynced on demand
5 tables · 8 relations detected
users
id
uuidPK
email
varchar
plan_id
uuidFK
status
enum
created_at
timestamp
subscriptions
id
uuidPK
user_id
uuidFK → users
plan
enum
expires_at
timestamp
status
varchar
invoices
id
uuidPK
sub_id
uuidFK → subs
amount
numeric
paid_at
timestamp
payment_methods
id
uuidPK
user_id
uuidFK → users
provider
varchar
last4
char(4)
credit_ledger
id
uuidPK
user_id
uuidFK → users
delta
integer
reason
varchar

Legend

PKPrimary Key
FKForeign Key

Auto-generated from
Prisma / Drizzle ORM

The Process

How it Works (~45s for Mid-Size Repos)

The 3-step process for codebase-led teams.

STEP 01

Connect Repo

Link any GitHub repository directly from your browser with one URL.

GitHub integration available
STEP 02

AI-Analysis

A fleet of AI agents parse and understand your codebase, database schemas, and relationships.

Understands your code in natural language
STEP 03

Get Docs, Diagrams & Chat

Get auto-generated API docs, interactive diagrams, schema visualizations, and chat with your codebase.

Interactive API reference, ER diagrams & AI chat
Security & Privacy

Built on trust, not just policy.

Three commitments — visible in how the product actually works, not just in a terms page.

01 — Permission Scope

Read-only GitHub permissions.

We request the minimum OAuth scope required — nothing more. DocFlow cannot push commits, create branches, or touch your repository settings.

oauth.scope = ["repo:read"]
Authorize DocFlow
OAuth 2.0
Read repository metadata
Granted
Read file contents & commit history
Granted
Write / push to branches
Blocked
Manage webhooks & settings
Blocked
Access private org data
Blocked
02 — Model Training

Your code is never used to train AI models.

Every AST parse and LLM query runs under enterprise zero-retention agreements. Your proprietary codebase never influences model weights.

training_data_retention = 0
Inference Pipeline
your_code.ts
source file
AST Parser
syntax tree extraction
LLM Inference
ephemeral · zero-retention
Enterprise Contract
/dev/null
source discarded — never stored
Model weights are never modified by customer data
03 — Data Lifecycle

Delete your indexed embeddings and repo data in 1 click.

Full data sovereignty. Disconnecting a repository triggers a hard cascade delete — vectors, AST cache, and all metadata purged immediately.

cascade_purge = immediate
acme-corp / backend-api
Indexed
Cascade delete scope
Vector embeddings
186 chunks
AST parse cache
48 files
Repo metadata
1 record
Chat history
3 sessions
Purge All Repository Data
The Features

Built for Developers & Businesses

Four focused tools that replace hours of manual work for technical teams and business stakeholders.

Auto API & Page Docs

Generates full OpenAPI 3.1 specs and human-readable docs from your controllers — no annotations required.

GET/api/usersList users · paginated200 OK
POST/api/reposConnect repository201
GET/api/repos/:id/schemaSynced on demand schema200 OK
POST/api/repos/:id/syncTrigger AST sync202
24 endpoints auto-documented from AST

Interactive DB Schema

Visualizes database relationships, FK constraints, and query paths synced on demand directly from your ORM models.

usersORM
idPK
emailvarchar
plan_idFK
subscriptionsORM
idPK
user_idFK → users
statusenum
users.id── 1:n ──subscriptions.user_id
Synced on demand
1:nFK6 models · 14 relations

Codebase RAG Chat

Ask questions about your code in plain English. DocFlow retrieves the exact file, line, and function — with full context.

Where is credit limit computed?
ME
DF

Found in src/policies/credit.ts:88

computeCreditLimit() — called by BillingService on every checkout event.

const limit = plan.base * multiplier;
Ask anything about your codebase…

AST Ingestion & Vector Indexing

Clones your repository, parses the Abstract Syntax Tree, and generates pgvector embeddings across all files, functions, and ORM models.

repo/ingestion-pipeline
COMPLETED
Git Tree Ingestion
48 files parsed
AST Route & API Extraction
24 APIs · 12 Pages
pgvector Code Embeddings
186 chunks indexed
PostgreSQL + Prisma AST100% indexed for RAG

Tested Languages & Frameworks

Verified across our core supported languages, backend frameworks, ORMs, and containerized infrastructure.

TypeScriptLanguage
JavaScriptLanguage
PythonLanguage
GoLanguage
RustLanguage
MarkdownDocs
YAMLConfig
Node.jsRuntime
TypeScriptLanguage
JavaScriptLanguage
PythonLanguage
GoLanguage
RustLanguage
MarkdownDocs
YAMLConfig
Node.jsRuntime
NestJSBackend
Next.jsFrontend
ReactFrontend
PrismaORM
PostgreSQLDatabase / pgvector
RedisCache / Queue
DockerDevOps
TailwindCSSStyling
NestJSBackend
Next.jsFrontend
ReactFrontend
PrismaORM
PostgreSQLDatabase / pgvector
RedisCache / Queue
DockerDevOps
TailwindCSSStyling

Enterprise-Grade Privacy: Your code never trains public models.

Built for Cross-Functional Teams

PRODUCT VELOCITY

Unblock Product Decisions

Product managers can validate technical feasibility, verify data structures, and inspect business logic boundaries without creating Jira tickets or scheduling engineer syncs.

  • Inspect production schema constraints without waiting for engineering syncs
  • Trace live user event triggers directly to telemetry tables
  • Verify feature flag dependencies and rollout stages in seconds
  • Export executive architecture summaries with one click
DocFlow Copilot Terminal
PM
Does the enterprise checkout endpoint support multi-currency tax exemptions?
DF

Yes. Referenced in `/services/billing/tax.ts:42`

Supports ISO-4217 currencies with automated VIES VAT verification.

{
"exemptionAllowed": true,
"supportedCurrencies": ["USD", "EUR", "GBP"],
"viesValidation": "strict"
}
Try it free

Start shipping better docs in seconds.

Turn any GitHub repository into a live, searchable doc system your entire team can trust from deployment to sprint.