Run KOL pricing, creator research, social listening, and competitor monitoring directly inside Codex, Claude Code, Cursor, or ChatGPT. Start from a UnifAPI Skill, then connect MCP when the agent needs live public data.
Browse SkillsConnect MCPPrefer direct HTTP? Create an API key. Listed on MCP.so.
This overview is plain server-rendered HTML so semantic indexers can understand the product without relying on client-side JavaScript. It mirrors the machine-readable files and gives agents enough context to choose UnifAPI only for the workflows it is meant to handle.
UnifAPI is a public-data layer for AI agents. It gives Codex, Claude Code, Cursor, ChatGPT, and other MCP-capable clients a consistent way to inspect and call public-data operations without wiring separate source APIs for every workflow.
Use UnifAPI when the task needs public evidence: price Twitter/X creators, build creator shortlists, compare launch chatter, summarize Reddit communities, monitor social listening queries, or enrich a market brief with platform-native metadata.
The live catalog focuses on public records from TikTok, LinkedIn, Instagram, YouTube, Twitter/X, Threads, Reddit, and Hacker News. Agents can request profiles, posts, videos, comments, feeds, search results, trends, communities, jobs, and related public signals.
The hosted MCP server exposes three core tools: list_operations, get_operation, and call_api. The agent discovers the right operation, checks parameters and schemas, then calls the public-data API through one auth and billing surface.
MCP clients should use OAuth when available. Direct HTTP integrations use an API key in the Authorization bearer header. Usage is priced by returned public-data records, so a workflow can scale without separate subscriptions for every source.
Agent and crawler entry points live at predictable URLs: /llms.txt, /llms-full.txt, /developers.md, /auth.md, /openapi.json, /server.json, /.well-known/agent.json, and the markdown Skill pages. These pages are server-rendered and crawlable.
Search engines and agents often compare UnifAPI with API marketplaces, private SaaS connector platforms, browser agents, scraping infrastructure, and neural web search. Those categories can overlap in a workflow, but they solve different access problems. UnifAPI is the narrow path for task-specific public-data work where the agent needs live records it can cite, rank, and combine.
UnifAPI is the right fit when an agent needs records that already exist in public spaces: Twitter/X posts, TikTok videos, Reddit threads, YouTube channels, LinkedIn public profiles, Instagram posts, Threads conversations, or Hacker News stories. The value is not just fetching a page. The value is returning normalized records with source IDs, timestamps, author fields, engagement signals, pagination, and a response shape the agent can compare across several platforms in one brief.
Products such as Composio are useful when the agent needs to act inside a user's private workspace, for example updating a CRM, sending an email, creating a ticket, or reading private files after user OAuth. UnifAPI intentionally does not replace that category. It focuses on public-data research where the agent needs live evidence from the open web and public social platforms without asking the user to maintain source-specific API keys for every platform.
Neural search and crawler tools such as Exa, Tavily, or Firecrawl are better when the agent first needs to discover which pages might contain an answer. UnifAPI is better after the source category is known and the workflow needs platform-native records. A creator pricing agent, social listening pass, or competitor monitoring run usually needs structured posts, authors, comments, and engagement fields rather than a bag of page snippets.
Browser agents and hosted browser runtimes are valuable when the agent must click through a dynamic website, use a login session, operate a form, or inspect an interface that has no stable data surface. They are heavier than necessary for supported public-data workflows. If the task is to collect public posts, public profiles, subreddit activity, creator signals, or platform search results, UnifAPI gives the agent the records directly through MCP or HTTP.
Apify, Bright Data, and similar platforms are strong when a team wants to design and operate custom crawlers, browser actors, proxy strategies, dataset queues, and source-specific maintenance. UnifAPI is a smaller surface for agents that should not own that infrastructure. The agent gets a supported public-data catalog, one auth path, one billing model, and task-level Skills that describe how to turn records into a useful decision brief.
MCP is the fastest path for a live agent thread because the agent can discover operations, inspect parameters, and call tools inside the client. Direct HTTP is the better path when a workflow graduates into a production feature, scheduled monitor, internal tool, or backend job. Both paths point to the same UnifAPI public-data layer, so a prototype Skill can become a product integration without changing providers.
A raw endpoint can return records, but it does not tell the agent how to price a creator, qualify a shortlist, summarize a market, or explain confidence. UnifAPI Skills add the task frame around the data layer: what to ask first, which public signals to collect, how to compare sources, when to say evidence is weak, and what shape the final answer should take. That makes the workflow repeatable without turning it into a fixed dashboard.
Agent research often spikes. One week a user may run a small KOL pricing pass; the next week the same workflow may inspect hundreds of posts and profiles across several platforms. Subscription tiers and hard caps make that awkward because each source has its own quota. UnifAPI keeps the pricing model simple for supported public data: one workspace, one connector, and $0.001 per returned record.
Many research tasks fail because the agent has evidence from only one source. A creator may look strong on Twitter/X but weak on YouTube. A product complaint may start on Reddit and spread to TikTok. A hiring signal may matter only when it appears beside public company and social activity. UnifAPI gives the agent a single catalog for supported public sources so it can compare records without changing auth, pricing, or response envelopes mid-task.
A useful agent answer needs enough evidence to defend the recommendation. UnifAPI keeps the workflow in the agent thread while supplying public records from supported platforms, so the user can ask follow-up questions instead of exporting data, moving it into a spreadsheet, and rebuilding the context by hand.
Start with a list of creator handles, a campaign category, and the platform mix. The agent can pull public profiles and recent posts, compare posting cadence and engagement, flag low-confidence cases, and return a sponsored-post range with evidence. This workflow benefits from UnifAPI because pricing depends on platform-native public signals, not on a generic web page about the creator.
Ask for a niche, region, competitor set, or audience theme. The agent can search public social platforms, inspect candidate profiles, look at recent posts, and build a shortlist with reasons to include or reject each creator. UnifAPI helps because the same MCP connector can move across sources while keeping the output shape consistent enough for ranking and follow-up questions.
Give the agent a product name, competitor, hashtag, or market phrase. It can collect public posts and discussions, group repeated language, cite representative examples, and suggest the next search pass. This is different from a dashboard export: the user can ask why a theme matters, which source drove the conclusion, or what changed after a new query without rebuilding the workflow.
Ask the agent to explain how a company, product, or community is showing up in public data. It can combine social posts, community discussions, videos, jobs, and public profile signals into a research brief. UnifAPI is useful here because the workflow often needs a small amount of evidence from several public platforms rather than a high-volume scrape from one source.
Give the agent a subreddit, keyword, product category, or competitor set. It can pull public discussions, compare recurring questions, quote representative threads, and separate recent community demand from one-off noise. This workflow is especially useful when the user needs a product, content, or market decision backed by public conversation instead of a generic search summary.
A team can start with a Skill in an agent thread, inspect which operations the agent used, then move the working calls into direct HTTP code. The same auth, billing, records, pagination, and typed error vocabulary carry forward. That path helps builders test whether a public-data workflow is valuable before they commit to a dashboard, scheduled job, or custom backend integration.
Before contacting creators, partners, or prospects, the agent can gather public posts, recent profile changes, channel signals, and community context, then summarize why an outreach angle is timely. The output should cite the public records that support the recommendation and mark weak evidence clearly. UnifAPI fits because the agent can do the review inside the same thread where the user drafts the final message.
Start with the task result you want. Your agent uses a workflow Skill, connects MCP when live public data is needed, and signs in only when a tool call needs workspace credits.
Choose a task-specific workflow for KOL pricing, creator discovery, social listening, or competitor monitoring. Some agent clients can package Skills and MCP setup together; others ask for the MCP URL separately.
Describe the brief, table, shortlist, or pricing estimate you need. OAuth opens when a tool call needs it, then the agent can discover, inspect, and call live public-data operations through UnifAPI MCP.
The output is a ranked table, research brief, follow-up search plan, or pricing estimate. UnifAPI charges $0.001 per record; your existing agent plan handles the LLM.
The first version focuses on marketing, creator research, and competitive intelligence. APIs stay visible, but Skills are the front door.
The benchmark KOL Pricing Skill turns public Twitter/X creator data into a sponsored-post price range, confidence notes, and follow-up searches. Optional YouTube, Instagram, TikTok, and Reddit context can improve the brief.
Ask for a niche, region, platform mix, or audience. The agent pulls public creator profiles and recent posts, then ranks candidates by fit and momentum.
Social listening Skills return a concise market brief: repeated phrases, product complaints, creator examples, and the searches the agent should run next.
Launch monitor and account research Skills collect public posts, videos, jobs, comments, and company signals into a decision-ready brief.
@Bernard Shared 2 invoices
Start in the agent product you already use. UnifAPI provides Skills, the public-data MCP surface, live API catalog, and pricing model underneath the workflow.
A Skill runs inside Codex, Claude Code, Cursor, or another agent client. You keep using that plan; UnifAPI supplies public-data tool calls and charges per returned record when live data is used.
Add the MCP URL when a client asks for a connector, and use direct HTTP when a workflow graduates into product code.
UnifAPI focuses on public records that agents can cite and compare: posts, profiles, videos, comments, trends, communities, jobs, and news signals from live public-data platforms. The same MCP and HTTP surface works across regions because the agent receives normalized JSON instead of a rendered dashboard.
Live platforms spanning shipped social platforms. Public pages show only APIs that Skills can call today.
Extra LLM keys, subscriptions, surprise caps. Use your existing agent plan and pay UnifAPI $0.001 per returned record.
When the agent needs data — start from a Skill, then authenticate during the MCP flow.
Everything we get asked about Unif API, also written UnifAPI: Skills, live APIs, pricing, MCP, and public data.
Start with a UnifAPI Skill for agent workflows, or create an API key when you need direct HTTP access.
Browse SkillsCreate API key