Connect AI Agent Tools with Locus Pro
Connect a personal agent or enterprise runtime to paid APIs through hosted MCP, choose the right credential, and verify the first tool call.
Learn how to connect agents, choose an integration, evaluate APIs, control spending, and bill customers with Locus Pro. Each guide links to the relevant implementation reference.
Connect a personal agent or enterprise runtime to paid APIs through hosted MCP, choose the right credential, and verify the first tool call.
Add an MCP server to Claude Code in two commands. Install the Locus plugin, sign in with /mcp, and let Claude call paid APIs from one credit balance.
Add Locus to the Code tab in the Claude desktop app. Add the plugin marketplace in Customize, sign in from a terminal, and fix a shadowing local entry.
Add Locus to Claude.ai or Cowork as a custom connector, upload three skills, and check your balance. Includes the Team plan Owner-only limit.
Add MCP to Codex: install the Locus plugin, run codex mcp login, and approve OAuth. Your agent gets paid APIs on one credit balance, no provider keys.
Add an MCP server to ChatGPT with Developer mode: create a Locus custom app with OAuth, select it in chat, and call paid APIs from one credit balance.
Add MCP to Cursor with a one-click install link or ~/.cursor/mcp.json. Approve Locus with OAuth and call paid APIs from one credit balance, no keys.
Set up Cline MCP for Locus Pro: install the remote server with OAuth, add the three Locus skills, and call paid APIs from one credit balance.
Add the Locus Pro MCP server to VS Code: save .vscode/mcp.json, start it from MCP: List Servers, approve OAuth, and call paid APIs from Copilot Chat.
Set up Gemini CLI MCP for Locus Pro: run gemini mcp add, sign in with /mcp auth, and let Gemini call paid APIs from one prepaid credit balance.
Add the Locus Pro MCP server to Grok. Set up a grok.com custom connector, install it in the Grok Build CLI with grok mcp add, or ask Grok Bot.
Add Locus to AI agents like OpenClaw, Hermes, Muse, and Instinct with one install prompt. See what it installs, what to approve, and what was tested.
How to connect an MCP server to an AI agent: the Locus Pro Streamable HTTP URL, remote MCP server OAuth with PKCE, the CLI, and headless bearer credentials.
Connect LangChain agents to the Locus Pro MCP server with MCPAdapter in Python or TypeScript, cap spend per call and fix errors. Brave search: 30.3 credits.
How LangGraph runs tools in a ToolNode, how to load MCP tools with MCPAdapter, and what a research run costs: at most 127.4 credits for 5 paid calls.
Where CrewAI tools come from, how to add MCP servers with MCPServerHTTP and a bearer header, and what a sales crew costs: 605 credits for 50 accounts.
Connect the OpenAI Agents SDK to MCP tools in Python or TypeScript, pick hosted or local, fix the 5-second timeout and cap spend. Search from 1.7 credits.
Add an HTTP MCP server to the Claude Agent SDK in TypeScript or Python, allow tools by name, fix needs-auth and cap spend. Stock quotes cost 5.1 credits.
Load remote MCP tools into generateText and streamText with createMCPClient from @ai-sdk/mcp, limit paid calls per request, and compare search costs.
Connect Pydantic AI to remote MCP tools with MCPToolset, force a max_credits ceiling in process_tool_call, pick error behavior. Email lookups: 1-10.1 credits.
Give n8n AI agents paid tools with the MCP Client Tool, HTTP Streamable and Bearer Auth, fix SSE-only nodes, and see costs: 2 credits per Apollo search.
Set up Mastra's MCPClient with a bearer header, choose static tools or per-request toolsets, gate paid calls, and price a 20-page scrape at 773 credits max.
Connect Google ADK agents to remote MCP tools with McpToolset and a bearer header, cap spend in before_tool_callback, fix timeouts. Searches up to 5.8 credits.
Set up OpenClaw step by step: install, onboard, add MCP servers, and connect Locus Pro so your agent can call paid tools from one balance.
Set up the Hermes agent step by step: install, first run, add MCP servers, and connect Locus Pro so your agent can call paid tools from one balance.
Find the best web search MCP server. Exa, Tavily, Brave, Firecrawl, Parallel and Perplexity compared on benchmarks, free tiers, setup and Locus Pro prices.
The best MCP servers by job: Context7, GitHub, Playwright, Chrome DevTools, Firecrawl, Exa and Supabase, with connect details, keys needed and data prices.
The best Cursor MCP servers: Context7, GitHub, Playwright, Chrome DevTools, Firecrawl and Exa, with mcp.json entries, which need a key, and marketplace status.
Codex MCP servers worth adding, with codex mcp add commands, config.toml options, OAuth login, plugins, and what replaced the removed codex mcp-server command.
ChatGPT MCP servers worth installing, from Exa to Alpha Vantage, plus the custom MCP server steps, local server tunnels, write confirmations and call prices.
Sales MCP servers compared: Apollo's official server, Hunter, HubSpot, why LinkedIn has none, and what a 100-lead list costs through Locus Pro.
SEO MCP servers compared: DataForSEO, Ahrefs from $129 a month, Semrush from $199, Search Console options, and pay-per-call SEO data through Locus Pro.
Finance MCP servers compared: Alpha Vantage's official server with a free key, SEC EDGAR servers, Yahoo Finance via yfinance, and market data through Locus Pro.
The most-installed Claude Code plugins, from frontend-design and superpowers to context7 and playwright, with install steps, context cost and data plugins.
The best Claude Code skills by install count, from grill-me and tdd to frontend-design and superpowers, with install commands and look-alikes to avoid.
Where to find MCP servers: the official MCP Registry, Glama's 118,530 listings, Smithery after its October 2026 shutdown, mcp.so, awesome lists and host stores.
Firecrawl's MCP server works with OAuth, an API key or no key. Its tools, setup for Claude Code and Codex, credits per call, plans, rate limits and Locus Pro.
Exa's MCP server at mcp.exa.ai works keyless for light use. Its four tools, setup for Claude Code and Codex, the free monthly credit, and calls via Locus Pro.
Perplexity's MCP server has search, ask, research and reason tools, needs an API key or admin OAuth, and bills at API rates. Setup, rate limits and options.
Tavily's MCP server has search, extract, crawl, map and research tools, with 1,000 free credits a month and a keyless mode. Setup, plans, limits and Locus Pro.
Apollo's MCP server works on any plan with OAuth and adds no fee. Credits per action, rate limits from 600 calls a day, plans from $49, and seat-free options.
LinkedIn has no official MCP server. How community servers drive your logged-in browser, what the User Agreement bans, and data APIs from 10.1 credits a call.
There is no official YouTube MCP server. Set up a community one with a Data API key, plan around 100 searches a day, and see what transcripts cost per call.
Reddit has no official MCP server, and new API access closed October 31, 2026. How community servers work, the 100 QPM limit, and per-call options from $0.0058.
Google has no web search MCP, and its Custom Search JSON API ends January 1, 2027. SerpApi and DataForSEO MCP setup, free tiers, and per-call Google results.
ElevenLabs' hosted MCP server signs in with OAuth and spends workspace credits, with 10,000 free a month. Its tools, catches, plans and per-call speech.
The most useful tools to install on a Hermes agent: web search, browser automation, paid APIs over MCP, and the hermes mcp commands to add each one.
The most useful tools to install on an OpenClaw agent: web search, browser control, paid APIs over MCP, and the openclaw mcp commands to add each one.
The best MCP servers for Claude Code: paid APIs over one connection, search, browser control, and repos, with the claude mcp add commands to install each.
YouTube's API only returns captions for videos you can edit. Compare youtube-transcript-api, Supadata, Scrape Creators and per-call access for agents.
Convert any web page to Markdown for an LLM. Jina Reader, Firecrawl, Cloudflare, Parallel, Tavily and Diffbot compared on free tiers, billing and catches.
Google closed the Custom Search JSON API to new users and ends it January 1, 2027. Compare Serper, SerpApi, Brave, Exa, Vertex AI Search and Gemini.
The SEC's EDGAR APIs are free JSON with no key, limited to 10 requests a second. Endpoints, User-Agent rules, and when SEC-API.io's $49 plan pays off.
Yahoo has no public Finance API, and yfinance is for personal use. Licensed options: Alpha Vantage, FMP and SEC EDGAR, with free tiers, limits and terms.
Zillow data needs Bridge approval, 1,000 calls a day and no storage. Compare the RentCast API, free Zillow CSVs and agent access with no RentCast plan.
Google shut QPX Express in 2018 and Amadeus self-service closed in 2026. Compare SerpApi's Google Flights engine, Duffel, Skyscanner and how each one charges.
Reddit's Data API is free to 100 queries a minute for approved apps, $0.24 per 1,000 calls above that. Access rules, AI terms and third-party options.
Gemini Deep Research runs $1 to $7 a task, Parallel $0.10 to $2.40 a run, Exa $0.012 to $1.00. Compare OpenAI, Perplexity, Tavily and capped options.
Track Google rankings by API: DataForSEO, Serper, SerpApi, SE Ranking and the free Search Console API, with how each bills and how many checks you need.
Firecrawl plans run from free (1,000 credits) to $749 a month for 1M credits. Credit costs per feature, worked examples, rate limits and the catches.
Tavily gives 1,000 free credits a month, then $0.008 a credit or plans from $30. Credits per search, extract and crawl, worked costs and catches.
Exa search costs $4 to $15 per 1,000 requests, contents $1 per 1,000 pages, and the free tier gives $10 a month. Rate limits, catches and examples.
SerpApi runs from 250 free searches to $725 for 100,000 a month. Price per 1,000 on every plan, speed-mode multipliers, renewal rules and examples.
Brave Search API costs $5 per 1,000 requests with $5 free each month; Answers adds $5 per million tokens. Rate limits, storage rights and examples.
Perplexity Search API costs $5 per 1,000 requests ($1 with Fast Search); the Agent API bills tokens plus tool fees. What changed for Sonar, with examples.
Apify is free with $5 of usage a month; paid plans cost $19, $199 and $999. Compute unit prices, pay-per-event Actors, worked costs and the catches.
DataForSEO is pay as you go with a $50 minimum. SERPs cost $0.0006 to $0.002, Labs $0.012 per task plus items. Multipliers, examples and catches.
Google Maps gives 10,000, 5,000 or 1,000 free events per SKU, then $2 to $40 per 1,000 for common SKUs. Field-mask rules, subscriptions and examples.
Clay pricing explained: Free, Launch from $167 and Growth from $446 a month, how actions and data credits work, rollover rules, and the cost of 1,000 leads.
ZoomInfo publishes no list prices. See the $33,500 median contract Vendr reports, ZoomInfo Lite, GTM AI API credits from $20, and the contract catches.
Apollo API pricing: which plans include the API, credits per endpoint, rate limits from 50 to 1,000 requests a minute, and what 1,000 enrichments cost.
Hunter.io pricing: Free, Starter $49, Growth $149 and Scale $299 a month, what a credit buys, the real cost per credit, API limits and agent access.
Email validation API prices compared: NeverBounce, ZeroBounce, Kickbox, Emailable and Hunter per check, unknown-result rules, expiry and 10,000-email costs.
People Data Labs pricing: Free up to 100 records, Pro from $98 a month for 350 person records, what a credit buys, API limits and agent access.
Prospeo pricing: Free with 100 credits, then $49, $99 and $249 per user a month. 1 credit per email, 10 per mobile, API rate limits and agent access.
What is Sumble? The tool that maps teams, technologies and hiring inside companies. Its Free and $99 Pro plans, API credit costs and agent access explained.
BuiltWith API pricing: $99 for 2,000 credits, plans from $295 a month, credits per endpoint, the free APIs, rate limits and the catches to check first.
Claude API rates from Haiku 5.5 at $0.10/$0.50 to Fable 5.2 at $8/$40 per million tokens, plus batch and caching discounts, rate limits and a cost example.
Gemini API rates per million tokens from $0.10 to $12, the free tier, Batch and Flex at 50% off, and the Argon and 3.8 Flash price rises due by January 2027.
DeepSeek API rates for deepseek-flash and V4 Pro at peak and off-peak, cache-hit prices, peak hours in UTC and US time, and a 10,000-call cost example.
xAI Grok API rates: Grok 4.7 at $2 input and $6 output per million tokens, long-context and batch rules, tool fees, rate limit tiers and a cost example.
Groq API prices: GPT-OSS 120B at $0.15/$0.60 per million tokens, Whisper from $0.04 an hour, free plan limits, 50% batch and cache discounts, and an example.
ElevenLabs API rates from $0.04 to $0.08 per 1,000 characters, Scribe v2 at $0.22 an hour, API plans from $5 to $825 a month, catches and cost examples.
Deepgram Nova-3 costs $0.0043 a minute pre-recorded and $0.0077 streaming, with $200 free. Add-ons, plans and Whisper API prices from OpenAI, Groq and others.
fal.ai prices per output: FLUX schnell at $0.003 per megapixel, Nano Banana from $0.0398 per image, video from $0.015 a second and H100 GPUs from $2.49 an hour.
The X (Twitter) API is pay per use: $0.005 per post read, $0.010 per user read, $0.015 per post created. Free credits, the 3M read cap and cost examples.
Compare API coverage, accounts and credentials, spend controls, billing ownership, data handling and upkeep when choosing Locus Pro or direct providers.
Evaluate agent tools with representative tasks, schema checks, useful results, latency, real charges, and failure handling before rollout.
Agentic payment platforms compared by job: prepaid API credits with Locus Pro, or cards and wallets from Skyfire, Nevermined, Crossmint, Stripe and Visa.
Which SERP API is cheapest at your volume? Scrapingdog, Serper, DataForSEO, Bright Data, SearchAPI, Brave and SerpApi compared on billing, free tiers and speed.
Firecrawl, Tavily and Apify alternatives for AI agents: what Exa, Parallel, Diffbot, Crawl4AI and Locus Pro do best, how each bills and their free tiers.
Clay alternatives for AI agents compared: Clay, Apollo, FullEnrich, Hunter, People Data Labs and Locus Pro, on setup, free tiers, billing and when to pick each.
Composio alternatives compared: Arcade, Pipedream Connect, Nango and Locus Pro on auth, catalog, billing and free tiers. Smithery is now a directory only.
Mapbox gives 5 to 10 times more free usage and lower rates per map load, geocode and route. Google’s mobile SDK is free, and its terms tie data to its map.
Apollo vs ZoomInfo compared: self-serve seats vs quoted yearly contracts, free tiers, what a credit buys, API and MCP access, renewal terms and which to choose.
Deepgram vs Whisper compared: streaming, speaker labels, file limits, hosted Whisper from $0.04 an hour, self-hosting and the 2027 whisper-1 shutdown.
ZoomInfo alternatives compared: Apollo, Lusha, UpLead, Cognism, RocketReach and per-call lookups for agents, on billing, free tiers, phones and fit.
Apollo.io alternatives by job: Prospeo and Lusha for data, Hunter and Snov.io for email and outreach, Kaspr for phones, plus Apollo per call for AI agents.
BuiltWith alternatives compared: Wappalyzer, Sumble, TheirStack and PredictLeads on data source, plan price, free tier and API access, plus per-call lookups.
ElevenLabs alternatives for text to speech: OpenAI, Deepgram, Cartesia, Google Cloud, Amazon Polly and xAI on billing, free tiers, licensing and fit.
RocketReach alternatives compared: ContactOut, Kaspr, Prospeo, Hunter, UpLead and People Data Labs on lookups, phones, free tiers and API access.
Clearbit no longer sells new standalone plans. Compare People Data Labs, The Companies API, Apollo, Logo.dev and Brandfetch for enrichment and logo APIs.
SerpApi alternatives compared on coverage and terms: Serper, SearchAPI, DataForSEO, Bright Data, Brave and Scrapingdog, with free tiers and legal cover.
Browserbase alternatives compared: Hyperbrowser, Steel, Browser Use, Anchor Browser, Kernel and Browserless on browser time, free tiers and a 50-hour example.
E2B alternatives for AI code sandboxes: Daytona, Modal, Vercel Sandbox, Cloudflare and Runloop on free credit, session limits, idle billing and fit.
RapidAPI is now Nokia API Hub. Compare Zyla API Hub, APILayer, Apify Store, Postman API Network and Locus Pro on catalogs, per-API billing and agent access.
Locus Pro vs AgentCash: card-funded credits or a USDC wallet paying x402 and MPP sellers, compared on free credit, spend limits, MCP setup and catalog.
Locus Pro vs Monid compared on catalog coverage, LLM access, sign-in, card or x402 payment, free credit, spend limits, BYOK and customer billing.
Locus Pro vs treg (treg.to) compared on catalog, sign-in, card payment, free credit, spend limits, team roles, BYOK, hosted MCP and self-hosting.
How to charge for AI usage: five billing methods, a worked cost per task, markup left after 2.9% + 30¢ card fees, and when to use prepaid credits.
AI credits explained: what GitHub, Salesforce, Lovable and Locus Pro credits are worth, how to price each action in credits, and when credits should expire.
Stripe, Metronome, Orb, Lago, OpenMeter and Stigg compared: published prices, a $50,000-a-month cost example and the 2025 to 2026 ownership changes.
How Stripe usage-based billing works in 2026: Meters API vs Metronome, 0.7% pricing, billing credits applied at invoice time, rate limits and setup steps.
How to monetize an API: five pricing models, RapidAPI's 25% fee vs Stripe's 0.7% plus card fees, and what you keep from $2,000 a month of API revenue.
Outcome-based pricing explained with Intercom's $0.99 per outcome, Zendesk and Sierra, plus how to price each result so it pays for failed attempts.
AI agent pricing models compared: seat, token, action, conversation, outcome and credits, with 2026 prices from Salesforce, Intercom, Claude and GitHub.
Track LLM cost per user: what to count, how Langfuse, Helicone and Locus Pro tag calls by customer, and a light, typical and heavy user margin check.
Can you resell API access? What OpenAI, Anthropic and Google Maps terms allow, how to price resold calls with a markup, and the risks a reseller carries.
Per-seat vs usage-based pricing for AI features: a worked 10-person team example, hybrid seat plus allowance plans, and what Claude, GitHub and Intercom charge.
Build vs buy for billing customers' agent API calls: what in-house metering takes, where it breaks, what you pay for, and when Stripe, Orb or Locus Pro fits.
Stripe usage-based billing vs Locus Pro for AI products: who supplies the APIs, how prepaid credits are enforced, costs, and when Stripe is the better pick.
Orb vs Metronome now that Adyen owns Orb and Stripe owns Metronome: prices, event fees, prepaid credits for AI agents, and when Locus Pro fits instead.
Choose hosted MCP for interactive agents or API and SDK access for application code. Compare discovery, credentials, billing, and customer attribution.
Understand Locus Pro credit balances, subscription allowances, purchased credits, usage pricing, reservations, refunds, and funding controls.
Design customer balances, exact usage prices, markup, retries, and receipts for an AI product, with a worked Locus Pro billing example.
Compare managed provider access with enterprise BYOK. Import supported OpenAPI actions, protect upstream credentials, and understand who pays.
Combine prepaid balances, account caps, scoped credentials, tool permissions, and live-quote ceilings to control agent tool costs.
Is there an OpenRouter for APIs? Yes: Locus Pro. Compare it with treg, Monid, AgentCash, TaskFuel, Orthogonal and more on billing, keys and MCP access.
What is x402? The open HTTP 402 payment standard explained: the payment flow, facilitators, EIP-3009, V2 changes, the Bazaar, and how Locus uses it.
What does HTTP 402 Payment Required mean? Why APIs return it, how to tell a credits error from an x402 or MPP challenge, and how a client should respond.
What is the Machine Payments Protocol (MPP)? How the Stripe and Tempo HTTP 402 standard works, which payment methods it supports, and how it compares to x402.
MCP is the open standard that lets AI apps call tools and read data. How hosts, clients and servers work, what the 2026-07-28 spec changed, and what calls cost.
MCP gateways compared: Docker, Kong, Cloudflare, agentgateway, AWS AgentCore and Locus Pro for paid APIs, with how each runs, how you pay and proxy rules.
How MCP server authentication works: the 401 challenge, protected resource metadata, Client ID Metadata Documents, PKCE, scopes and a real handshake.
Remote vs local MCP servers: how Streamable HTTP works in the 2026-07-28 spec, how to add one with OAuth, Claude Code output limits, and per-call costs.
MCP security risks explained: tool poisoning, prompt injection, confused deputy, token passthrough and runaway spend, with the spec rules and a checklist.
Agent skills are folders with a SKILL.md file that agents load on demand. The open format, the 100-token cost, where Claude Skills run, and safety checks.
Skills vs MCP: skills teach an agent how to do a task, MCP gives it tools and data. Compare context cost, credentials, network access and when to use both.
AI agent governance in practice: identity, least privilege, tool allowlists, spend limits, approvals and audit logs, with Entra, Okta, IMDA and AWS examples.
AI agent observability compared: Langfuse, LangSmith and Datadog free tiers and prices, a 10,000-run cost example, and how to track paid tool spend.
How Claude Code searches the web: WebSearch and WebFetch limits, the 200-search session cap and how to raise it, and adding Brave or Perplexity through MCP.
Pay x402 and MPP APIs from Claude Code, Cursor or Codex. Compare Locus Pro credits with AgentCash, PayAPI Market, Bazaar MCP and paying sellers yourself.
Learn how AI agents pay for tools and APIs from a prepaid balance under predefined limits, price ceilings, and audit controls, over MCP, HTTP, and x402.
For current offerings, see the tool catalog, pricing, and enterprise overview.