API for Content Style Refinement at Agency Scale
CitationBench gives agencies a dedicated API for content style refinement with reusable refiners that enforce brand voice and quality standards automatically across every client workspace.
Brand Voice Breaks at Scale
Managing distinct brand voices across dozens of clients leads to inconsistent output and manual rework that kills margins.
Every client sounds the same
Generic AI writing tools ignore client-specific tone guidelines, producing content that blurs brand identity across your portfolio.
Manual QA bottlenecks growth
Editors spend hours re-polishing drafts to match brand voice, capping how many clients your team can realistically serve.
Style rules live in documents nobody reads
Brand guidelines stored in Google Docs or Notion never make it into the actual content production workflow consistently.
The fix
Reusable Refiners. One API. Every Client.
CitationBench lets you configure reusable content refiners once per account and apply them automatically across all client workspaces via REST API or MCP. Brand voice, CTA rules, FAQ injection, and competitor removal run at publish time without manual editing.
Every Tool Your Content Pipeline Needs
From first-draft generation through style enforcement, evaluation, and publishing, CitationBench covers the full production workflow in a single API surface.
Style API
Reusable Brand Voice Refiners Per Client
Configure brand voice, CTA rules, FAQ injection, TLDR summaries, and competitor removal once via `produce.refine.*`. Refiners apply automatically across all content in a client workspace without a manual step.
- Enforce tone per workspace
- Auto-apply at publish time
- Bulk refine existing content
Quality Control
Score Every Draft Before It Ships
Use `produce.evaluate.score` to grade every draft on readability, SEO quality, and brand alignment. Build custom rubrics per client and compare variants with `produce.evaluate.compare` before any content goes live.
- Custom evaluation rubrics
- A/B content comparison
- Brand alignment scoring
Content Production
Long Form Drafts Ready for Refinement
Generate long-form SEO drafts via `produce.blog_post.create` or in bulk with `produce.blog_post.bulk_create`. Every draft feeds directly into the content refiner pipeline so style rules apply before the piece ever reaches a human editor.
- Bulk create from keyword lists
- Section-level regeneration
- Brief to refined output pipeline
Publishing
Publish Refined Content Across Every CMS
Send refined, evaluated content to WordPress, Ghost, Wisp, or any custom endpoint via `produce.publish.send`. Platform credentials are isolated per client workspace so one API key manages your entire portfolio cleanly.
- WordPress and Ghost supported
- Custom REST webhooks
- Per-workspace credentials
Agency Architecture
One API Key For Every Client Brand
Manage N client workspaces under a single master API key. Every refiner config, content draft, and evaluation rubric is scoped to the correct client via the `X-Workspace-Id` header. White-label-ready architecture built for agencies running 10 to 100 brands.
- X-Workspace-Id scoping
- Isolated style configs
- Bulk actions across workspaces
How it works
From sign-up to first call in five minutes.
Sign up and get your key
An `sk_test_*` key lands in your dashboard instantly. No demo gate — start calling real endpoints with shape-complete responses.
Add the MCP server or hit REST
`claude mcp add citationbench https://mcp.citationbench.com/mcp` from Claude Code, Cursor, or any MCP client. Or `curl` against `api.citationbench.com/v1/*` directly.
Run a tool or invoke an agent
Every tool returns shape-complete demo data without auth, so your agent works before the user signs up. Tools across research, production, indexing, link-building, and agents.
Scope to a workspace and ship
Add `X-Workspace-Id: ws_***` to scope per client. Switch to a live `sk_live_*` key when you're ready. Same API surface, same SDK.
Why CitationBench
Built for production, agency-scale, and AI-agent-first.
~35 tools, one MCP server
Hosted at mcp.citationbench.com/mcp. Works with Claude Code, Cursor, Claude Desktop, Windsurf, and ChatGPT Apps. No self-hosting required.
Durable jobs, not fire-and-forget
Every long-running call runs on Cyclonic workers — survives restarts, cancellable, resumable, streams via SSE. Production-safe by default.
Multi-workspace from day one
One master API key, N client workspaces. Switch with a single `X-Workspace-Id` header. Per-client data isolation, bulk ops across all of them.
Demo mode out of the box
Every endpoint responds in shape-complete demo mode without auth, so you can build the agent before the user signs up.
FAQ
Common questions
Start Enforcing Brand Voice at Scale
Configure reusable content refiners once and apply consistent brand voice across every client automatically via REST API or MCP.
Start Free