Niche for developers

Build agents on the
editorial intelligence MCP.

25 workflow-grained MCP tools, signal to angle to draft to render to publish, with verifier-checked claims and source provenance on every output. Install on Claude Desktop, Claude Code, Cursor, or any MCP-compatible client in 5 minutes.

Five-minute install

1. Generate a PAT at /settings/api-keys. Tokens prefix with niche_sk_.

2. Add the server to your agent client. Claude Code (CLI):

claude mcp add --transport http niche https://api.nicheangle.com/mcp \
  --header "Authorization: Bearer niche_sk_..."

Claude Code (project .mcp.json) — note the field is type, value http:

{
  "mcpServers": {
    "niche": {
      "type": "http",
      "url": "https://api.nicheangle.com/mcp",
      "headers": {
        "Authorization": "Bearer niche_sk_..."
      }
    }
  }
}

Claude Desktop: Settings → Connectors → Add custom connector → paste the URL, choose Streamable HTTP, add your PAT. Cursor (~/.cursor/mcp.json): just url + headers (no transport field).

3. Restart the client (or open a new thread) — MCP clients cache the connection (auth + the tools/list) at connect time, so reconnect after changing your token or to pick up new tools. Your agent then sees all 25 Niche tools.

The 25-tool surface

Tools group into 7 functional bands. Full schemas live at /.well-known/agent-skills/index.json (auto-generated from the live registry, always in sync with the server's tools/list response).

Discovery + state

7 TOOLS

Start an editorial run for a niche, or ask an analyst-shaped intelligence query (ranked slate + engine-grounded, fact-verified narratives). Poll or long-poll session state with `since_status`, list past sessions, revert to any prior checkpoint, cancel mid-flight, export a finished run.

niche_signal_scanniche_intelligence_queryniche_session_stateniche_list_sessionsniche_session_revertniche_session_cancelniche_session_export

Editorial

1 TOOL

Propose 3 frame-aware angles per picked story with frame, hook, tension, CTA, a verifier audit, and a recommended_angle_id scored against the brand profile. Pass custom_framing to author your own angle.

niche_angle_propose

Drafting

5 TOOLS

Generate platform-native drafts with trust fields (faithfulness score + claim count, ungrounded claims, source diversity, recency). Sub-field revision with a diff[] return — you supply the new text. Add new output cells after review, or reuse an existing rendered image across cells. Or skip research and draft straight from a take or URL (niche_draft_direct).

niche_draft_directniche_draft_createniche_draft_reviseniche_add_outputniche_reuse_asset

Rendering

3 TOOLS

On-demand image card and 9:16 reel with cell-aware dimensions (1200×627 for linkedin:image_post, 1080×1350 for instagram:image_post, etc.). Idempotent replace. Or attach the creator's own photo as the post image (niche_attach_image).

niche_render_image_cardniche_render_reelniche_attach_image

Brand intelligence

7 TOOLS

Persisted BrandProfile JSON (voice rules, lexicon, framing allowlist, verifier overrides). Ingest brand from URL, files, or freeform text (async — poll niche_brand_kit_ingest_status). Voice extraction from past posts. Guided-setup question chain for first-touch onboarding.

niche_brand_profile_getniche_brand_profile_setniche_brand_kit_ingestniche_brand_kit_ingest_statusniche_brand_kit_updateniche_brand_kit_guided_setupniche_voice_profile_ingest

Publishing

1 TOOL

Dry-run by default. Idempotency-key required for commit. Verifier-blocked outputs refuse to publish even at `dry_run=false`.

niche_draft_publish

Account + orientation

1 TOOL

One read-only call (niche_whoami) that returns the authenticated account, plan, and credit balance; the server version + endpoint + live tool count; the capability map (every tool grouped by band + the recommended flow); and the brand state, so an agent knows before a run whether a brand profile / kit / voice will shape the output.

niche_whoami

Trust fields on every draft

Every tool response carries its sources, an ungrounded-claim flag, and source-diversity + recency checks — provenance on every output, the only editorial MCP that ships receipts. Agents can gate publish decisions on the trust block without re-deriving it:

verifier_blocked_reason + verifier_blocked_claim: present when a load-bearing fabrication couldn't be grounded. Publish refuses even at dry_run=false.
source_faithfulness_score ∈ [0.0, 1.0]: fraction of concrete claims in the draft that appear verbatim or normalized in the source material.
source_ungrounded_claims: list of specific tokens (numbers, dates, proper nouns) the verifier couldn’t ground. Agents typically gate at >0 entries.
source_diversity_passed + source_recency_passed: boolean checks on the contributing signal pool.

Endpoints

MCPapi.nicheangle.com/mcp · Streamable HTTP · Bearer PATOpenAPIapi.nicheangle.com/openapi.jsonAgent skills/.well-known/agent-skills/index.jsonAPI catalog/.well-known/api-catalog (RFC 9727)

Going deeper

The full walkthrough, covering brand-kit setup, voice profile, a sample end-to-end agent run, and troubleshooting, lives at /integrations. That page covers both the human web app and the agent surface; this page is the developer-shaped subset.