"An MCP for interacting with Google Chrome DevTools and aiding in scraping" matching MCP tools:

• onyx_solve_captcha

  onyx-mcp

  Solve an image-based text captcha and return the recognized text. Works on standard alphanumeric captchas (web signup forms, login walls, scraping checkpoints). OCR via ddddocr — typical p50 latency 30-80ms, 70-90% accuracy on common captcha fonts. Provide either an image URL we fetch on your behalf, or raw base64 image bytes if you already have them. Use when an agent encounters a captcha mid-task and needs to continue without human intervention. Cheaper and faster than 2captcha for simple image captchas; not designed for reCAPTCHA v2/v3 or hCaptcha (those are interaction-based). (price: $0.003 USDC, tier: metered)

  Connector

• tronsave_internal_create_extend_request

  Tronsave MCP Testnet

  Submit an extension request for existing delegated resources on TronSave via authenticated REST `POST /v2/extend-request`. Requires a logged-in MCP session created by the `tronsave_login` tool: include `mcp-session-id: <sessionId>` returned by `tronsave_login` on subsequent MCP requests. Internal tools never accept API keys via tool arguments; the server forwards the API key cached in session to TronSave internal REST endpoints. Side effect: creates an extension order and may commit TRX from the internal account. `extendData` must follow the REST contract (see schema on each row). Populate it from TronSave outside this MCP—for example the authenticated `POST /v2/get-extendable-delegates` response field `extendData`, or another TronSave client. Do not copy rows blindly from `tronsave_list_extendable_delegates` (GraphQL); that payload shape differs and is for market discovery only.

  Connector

• ceevee_list_versions

  CareerProof MCP

  List all CV versions for the authenticated user. Returns an array of version objects with id, filename, created_at, and main_version flag. Use the version id as cv_version_id in ceevee_analyze_positioning, ceevee_full_review, ceevee_confirm_lens, and ceevee_chat. Free.

  Connector

• get_user_profile

  anamnese

  PRIMARY TOOL - Call this at the START of every conversation to load comprehensive user context. Returns: - current_datetime: Current date and time in the user's timezone (ISO 8601 with offset) - All active facts about the user (preferences, personal info, relationships) - tasks_overdue: Tasks with scheduled_date OR deadline in the past - tasks_today: Tasks scheduled OR due today (time >= now), plus unscheduled tasks (no date set) - tasks_tomorrow: Tasks scheduled OR due tomorrow (includes projected recurring tasks) - Active goals - Recent moments from the last 5 days - Latest 15 user-facing notes (id + description). Use get_note to retrieve full content. - ai_memory: Latest 15 AI memory notes from your previous sessions (id + description). Use get_note to retrieve full content. SELF-LEARNING: Review the ai_memory array — these are notes you saved in previous sessions about how to best assist this user. Load relevant ones with get_note. Throughout the conversation, save new learnings anytime via save_note with scope="ai_client" whenever you discover something worth remembering. - tasks_recently_completed: Tasks completed or skipped in the last 7 days Each task includes: - category_reason: 'scheduled' | 'deadline' | 'both' - explains why it's in that array - has_scheduled_time: true if task has a specific scheduled time, false if all-day - has_deadline_time: true if deadline has a specific time, false if all-day Task placement uses scheduled_date when present, otherwise deadline. Each task appears in exactly one category. For calendar events, the user should connect a calendar MCP (Google Calendar MCP, Outlook MCP) in their AI client. Query those MCPs alongside Anamnese for a complete daily view. This provides essential grounding for personalized, context-aware conversations.

  Connector

• tronsave_login

  Tronsave MCP Testnet

  Create an authenticated server session and return a `sessionId` for subsequent tool calls. Default mode is wallet signature login for platform tools; secondary mode is `apiKey` login for internal tools. For wallet login, ALWAYS call `tronsave_get_sign_message` first, sign that exact message client-side, then call `tronsave_login` with `signature_timestamp` in exact format `<signature>_<timestamp>` (signature and timestamp joined by `_`). Use returned `sessionId` as `mcp-session-id` on every subsequent request.

  Connector

• frogeye_batch_scan

  Frogeye Security Scanner

  Batch scan up to 10 code snippets in a single MCP call. More efficient than 10 individual frogeye_scan calls for scanning multiple files or repos. Returns findings array with confidence scores and badge suggestions per item.

  Connector

• atlas_list_jds

  CareerProof MCP

  List all job descriptions for a hiring context. Returns an array of JD objects with id, title, and content. Use JD content as jd_text in atlas_fit_match, atlas_fit_rank, and atlas_start_jd_fit_batch. Requires context_id from atlas_create_context or atlas_list_contexts. Free.

  Connector

• get_usage_stats

  Name Whisper — ENS Intelligence Layer

  Get usage statistics for this MCP server session. Returns tool call counts, success rates, and average latency.

  Connector

• thinkneo_bridge_a2a_to_mcp

  ThinkNEO Control Plane

  Bridge an A2A (Agent-to-Agent Protocol) task to an MCP server. Receives an A2A task, identifies the best matching MCP tool on the target server, executes it, and returns the result wrapped in A2A response format. Enables A2A agents to use any MCP server transparently. Extracts the intent from the A2A task, maps it to an MCP tool, calls the tool, and wraps the result in A2A response format. Use this to let A2A agents interact with any MCP server. Requires authentication.

  Connector

• thinkneo_bridge_mcp_to_a2a

  ThinkNEO Control Plane

  Bridge an MCP tool call to an A2A (Agent-to-Agent Protocol) agent. Maps MCP tool name and parameters to the A2A task format, enabling interoperability between MCP servers and A2A agents. Returns a ready-to-send A2A task object with full protocol compliance. Translates the MCP tool_name and arguments into an A2A task, sends it to the target A2A agent, waits for completion, and translates the response back to MCP format. Use this to make any MCP tool accessible to A2A agents (Google's agent ecosystem). Requires authentication.

  Connector

• teams.list

  Koalr

  List all engineering teams in the organization with their member counts and slugs. Use this to discover team IDs needed for filtering other metrics tools. Returns an array of team objects with id, name, slug, and memberCount. Read-only.

  Connector

• atlas_list_custom_eval_models

  CareerProof MCP

  List all custom evaluation models for the authenticated user. Returns an array of model objects with id, name, description, and status. Use model id in artifact, rubric, and evaluation tools. Free.

  Connector

• strale_web3_assurance

  Strale - 290+ API capabilities for AI agents

  Returns a decision-ready answer about an on-chain counterparty (wallet, smart contract, token, DeFi protocol, or bridge) in a single call. Surfaces verdict (proceed/review/block/insufficient_evidence), reason_codes (machine-parsable UPPERCASE_SNAKE_CASE), critical_flags, suggested_action, evidence map (sanctions, mixer-graded, scam-cluster, wallet-history, token-safety, contract-verification, protocol-risk, EAS attestations, ERC-8004 reputation, more), and a sidecar audit_url. Two modes: 'outbound' (agent vetting recipient pre-payment, full evaluator set, 8s budget) or 'reverse-call' (service publisher gating an inbound x402 buyer in real-time, critical evaluators only, sub-second SLA). Use before any agent transacts on-chain — sending value, swapping, staking, minting, bridging, or interacting with a contract.

  Connector

• list_session_slides

  Slidev MCP

  List all slide presentations created in the current MCP session. Returns URLs, themes, and timestamps for each presentation you've created.

  Connector

• fetch-page

  PartsTable

  Fetch a web page and return its content as text, Markdown, or HTML. Includes rate limiting (2s per domain, max 10 req/min) for legal compliance. Automatically handles HTML-to-text conversion. Max response size: 1MB. Use for OEM verification and manufacturer website scraping.

  Connector

• images_prepare_upload

  Roboflow (Official)

  Get an upload URL to upload a single image to a project. Returns a pre-built upload URL and instructions. The caller must perform the actual upload using curl since the MCP server cannot access local files. This endpoint uploads images only. To add annotations, call annotations_save with the image ID from the upload response. For bulk uploads with annotations, use images_prepare_upload_zip.

  Connector

• hasdata_google_serp_immersive_product_getImmersive_e29f691177

  HasData

  Get Immersive Product Information Expands the Google Shopping Immersive Product pop-up given an immersiveProductPageToken from the Google Shopping API, with optional moreStores (up to ~13 merchants instead of 3–5) and nextPageToken for paginating stores. Returns multi-store offers (merchant, price, shipping, condition, URL), product specs, images, ratings, and the nextPageToken. Use for price-comparison bots, merchant discovery, dropshipping research, and aggregating full offer lists per product.

  Connector

• notes_save

  DialogBrain

  Save a note to your notebook. In agent mode writes to your own notebook by default; agents cannot write to other agents' notebooks. In MCP mode target_agent_id is required. If a note with the same key and scope already exists, it will be updated. Use scope to organize: 'global' for general knowledge, 'thread' for thread-specific context, 'person' for contact-specific info.

  Connector

• bing_search

  Searchapi

  Search the web using Bing. Returns organic results, related searches and more. Alternative to Google for web search with different ranking algorithms and results.

  Connector

• delete_endpoint

  Singularity Layer MCP

  ⚠️ DESTRUCTIVE: Permanently delete an endpoint. Cannot be undone. Only works for agent-created endpoints. PATs are preferred and require mcp:endpoints:write or mcp:*.

  Connector