Niche discovery

Discovery watches the platforms your audience uses for content in your niche. You describe the Brand in plain language; Sigmora compiles that into search terms per platform, collects matching content on a schedule, and shows you what is performing, with the evidence and freshness of each result.

Create a Brand

Discovery belongs to a Brand. Its niche and description are what Sigmora uses to decide what to watch, so make them specific. You do not need a theme id: creating the Brand detects its theme profile from this text.

curl -X POST https://api.sigmora.org/v1/brands \
  -H "Authorization: Bearer $SIGMORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Acme Money",
    "niche": "personal finance for first-time investors",
    "description": "Practical saving, credit, and investing education for young professionals"
  }'

import os
import requests

session = requests.Session()
session.headers["Authorization"] = f"Bearer {os.environ['SIGMORA_API_KEY']}"

brand = session.post(
    "https://api.sigmora.org/v1/brands",
    json={
        "name": "Acme Money",
        "niche": "personal finance for first-time investors",
        "description": "Practical saving, credit, and investing education for young professionals",
    },
    timeout=30,
).json()["brand"]
brand_id = brand["id"]

const headers = {
  Authorization: `Bearer ${process.env.SIGMORA_API_KEY}`,
  "Content-Type": "application/json",
};

const { brand } = await fetch("https://api.sigmora.org/v1/brands", {
  method: "POST",
  headers,
  body: JSON.stringify({
    name: "Acme Money",
    niche: "personal finance for first-time investors",
    description: "Practical saving, credit, and investing education for young professionals",
  }),
}).then((r) => r.json());
const brandId = brand.id;

Save a discovery plan

PUT /v1/discovery/configs/{brandId} compiles the Brand’s profile into a watch plan and saves it. Choose the platforms, a region and how often to collect.

curl -X PUT https://api.sigmora.org/v1/discovery/configs/brd_4b81 \
  -H "Authorization: Bearer $SIGMORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "enabled": true,
    "platforms": ["youtube_shorts", "tiktok", "reddit", "news"],
    "regionCode": "US",
    "discoveryIntervalMinutes": 360,
    "refreshIntervalMinutes": 60
  }'

plan = session.put(
    f"https://api.sigmora.org/v1/discovery/configs/{brand_id}",
    json={
        "enabled": True,
        "platforms": ["youtube_shorts", "tiktok", "reddit", "news"],
        "regionCode": "US",
        "discoveryIntervalMinutes": 360,
        "refreshIntervalMinutes": 60,
    },
    timeout=30,
).json()
print(plan["config"], plan["health"])

const plan = await fetch(`https://api.sigmora.org/v1/discovery/configs/${brandId}`, {
  method: "PUT",
  headers,
  body: JSON.stringify({
    enabled: true,
    platforms: ["youtube_shorts", "tiktok", "reddit", "news"],
    regionCode: "US",
    discoveryIntervalMinutes: 360,
    refreshIntervalMinutes: 60,
  }),
}).then((r) => r.json());
console.log(plan.config, plan.health);

Platforms are youtube_shorts, instagram_reels, facebook_reels, tiktok, twitter, reddit, news, medium, google_news and google_web. Add intent, keywords and excludeKeywords to steer the plan; preview them first with POST /v1/discovery/preview (see Content intelligence). A new plan is saved with enabled: false unless you turn it on, so saving never starts automation by surprise. The saved queryPlan records exactly what is watched:

the primary and supporting theme ids;
seed terms from the Brand’s niche, description and theme path;
the search terms used on each platform;
what each query looks for — momentum, news, controversies, audience questions or competitors;
the compiler version and when it ran.

The plan is stored rather than regenerated on every run, so results are reproducible and you can see what Sigmora was watching.

Check source health

curl https://api.sigmora.org/v1/discovery/configs/brd_4b81/health \
  -H "Authorization: Bearer $SIGMORA_API_KEY"

Health reports which platforms can be searched right now, plus the last discovery and metric refresh. You never send platform credentials. A source reported as missing_credentials is not available yet — choose a source marked ready, or contact Sigmora support. Each source also reports a queryMode: provider-query means your watch terms are sent to that platform’s search; broad-ingest-theme-filter means a broader Reddit or news feed is read and filtered to your themes. ready means the source can be queried, not that the platform has full coverage or unlimited quota.

Run discovery now

Enabled plans run on their schedule. Scheduled discovery runs at most every 6 hours and scheduled refreshes at most every 30 minutes, so a discoveryIntervalMinutes below 360 or a refreshIntervalMinutes below 30 runs at that floor. Scheduled runs are also bounded per UTC day across all plans: by default 32 discoveries and 96 refreshes a day (plans on the same primary theme share one refresh). The bound is spread over the day, so each pass takes an even share of what is left: 8 discoveries every 6 hours and 2 refreshes every 30 minutes. Up to 4 themes keep the default 60-minute refreshIntervalMinutes, and up to 24 themes are each refreshed at least every 6 hours. Past that, plans take turns: a plan that waits stays due and goes first at a later pass. To run one immediately:

curl -X POST https://api.sigmora.org/v1/discovery/runs \
  -H "Authorization: Bearer $SIGMORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contentBrandId": "brd_4b81", "kind": "discover"}'

run = session.post(
    "https://api.sigmora.org/v1/discovery/runs",
    json={"contentBrandId": brand_id, "kind": "discover"},
    timeout=120,
).json()["run"]

const { run } = await fetch("https://api.sigmora.org/v1/discovery/runs", {
  method: "POST",
  headers,
  body: JSON.stringify({ contentBrandId: brandId, kind: "discover" }),
}).then((r) => r.json());

Use kind: "refresh" to re-check metrics for content already found. Refresh is richest for YouTube, where each check is stored as a timestamped snapshot, so results can report view velocity once there are two observations. A run across several sources can finish as partial: results from the sources that succeeded are kept along with the errors from those that did not. One provider failing never erases what the others found. List and inspect runs, including the exact query plan each one used:

GET /v1/discovery/runs?contentBrandId=brd_4b81
GET /v1/discovery/runs/{runId}

The plan’s lastDiscoveryAt and lastRefreshAt show whether scheduled collection is moving.

See what is performing in the niche

GET /v1/discovery/configs/{brandId}/corpus returns the collected content that matches the Brand’s watch terms (or its niche, if it has no plan yet), with the terms used and the health of each source.

curl "https://api.sigmora.org/v1/discovery/configs/brd_4b81/corpus?sort=top&platform=tiktok" \
  -H "Authorization: Bearer $SIGMORA_API_KEY"

sort=top ranks by score relative to the niche; sort=newest by collection time. platform filters to one source. Because sources is included, an empty items list tells you why — for example, a platform that is not available yet. A Brand with a niche but no plan gets the default plan the first time it is read, as in the dashboard. A Brand with neither has no niche to read: items is empty and note says so. Reading the corpus is free. Over MCP it is discovery_corpus_get.

Get started

Create

Grow & automate

Integrate

Build reliably

Operate

Create a Brand

Save a discovery plan

Check source health

Run discovery now

See what is performing in the niche

Next steps

Themes and trends

Content autopilot

​Create a Brand

​Save a discovery plan

​Check source health

​Run discovery now

​See what is performing in the niche

​Next steps

Themes and trends

Content autopilot

Create a Brand

Save a discovery plan

Check source health

Run discovery now

See what is performing in the niche

Next steps