What TinyFish is Shipping<\/strong><\/h3>\nHere is what each product does:<\/p>\n
\n- Web Agent<\/strong> \u2014 Executes autonomous multi-step workflows on real websites. The agent navigates sites, fills forms, clicks through flows, and returns structured results without requiring manually scripted steps.<\/li>\n
- Web Search<\/strong> \u2014 Returns structured search results as clean JSON using a custom Chromium engine, with a P50 latency of approximately 488ms. Competitors in this space average over 2,800ms for the same operation.<\/li>\n
- Web Browser<\/strong> \u2014 Provides managed stealth Chrome sessions via the Chrome DevTools Protocol (CDP), with a sub-250ms cold start. Competitors typically take 5\u201310 seconds. The browser includes 28 anti-bot mechanisms built at the C++ level \u2014 not via JavaScript injection, which is the more common and more detectable approach.<\/li>\n
- Web Fetch<\/strong> \u2014 Converts any URL into clean Markdown, HTML, or JSON with full browser rendering. Unlike the native fetch tools built into many AI coding agents, TinyFish Fetch strips irrelevant markup \u2014 CSS, scripts, navigation, ads, footers \u2014 and returns only the content the agent needs.<\/li>\n<\/ul>\n
The Token Problem in Agent Pipelines<\/strong><\/h3>\nOne of the consistent performance problems in agent pipelines is context window pollution. When an AI agent uses a standard web fetch tool, it typically pulls the entire page \u2014 including thousands of tokens of navigation elements, ad code, and boilerplate markup \u2014 and puts all of it into the model\u2019s context window before reaching the actual content.<\/p>\n
TinyFish<\/a> Fetch addresses this by rendering the page in a full browser and returning only the clean text content as Markdown or JSON. The company\u2019s benchmarks show CLI-based operations using approximately 100 tokens per operation versus roughly 1,500 tokens when routing the same workflow over MCP \u2014 an 87% reduction per operation.<\/p>\nBeyond token count, there is an architectural difference worth understanding: MCP operations return output directly into the agent\u2019s context window. The TinyFish CLI writes output to the filesystem, and the agent reads only what it needs. This keeps the context window clean across multi-step tasks and enables composability through native Unix pipes and redirects \u2014 something that is not possible with sequential MCP round-trips.<\/p>\n
On complex multi-step tasks, TinyFish reports 2\u00d7 higher task completion rates using CLI + Skills compared to MCP-based execution.<\/p>\n
The CLI and Agent Skill System<\/strong><\/h3>\nTinyFish<\/a> is shipping two developer-facing components alongside the API endpoints.<\/p>\nThe CLI<\/strong> installs with a single command:<\/strong><\/p>\n\n\n\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nCopy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\nnpm install -g @tiny-fish\/cli<\/code><\/pre>\n<\/div>\n<\/div>\nThis gives terminal access to all four endpoints \u2014 Search, Fetch, Browser, and Agent \u2014 directly from the command line.<\/p>\n
The Agent Skill<\/strong> is a markdown instruction file (SKILL.md) that teaches AI coding agents \u2014 including Claude Code, Cursor, Codex, OpenClaw, and OpenCode \u2014 how to use the CLI. Install it with<\/strong>:<\/p>\n\n\n\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nCopy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\nnpx skills add https:\/\/github.com\/tinyfish-io\/skills --skill tinyfish<\/code><\/pre>\n<\/div>\n<\/div>\nOnce installed, the agent learns when and how to call each TinyFish endpoint without manual SDK integration or configuration. A developer can ask their coding agent to \u201cget competitor pricing from these five sites,\u201d and the agent autonomously recognizes the TinyFish skill, calls the appropriate CLI commands, and writes structured output to the filesystem \u2014 without the developer writing integration code.<\/p>\n
The company also notes that MCP remains supported. The positioning is that MCP is suited for discovery, while CLI + Skills is the recommended path for heavy-duty, multi-step web execution.<\/p>\n
Why a Unified Stack?<\/strong><\/h3>\nTinyFish<\/a> built Search, Fetch, Browser, and Agent entirely in-house. This is a meaningful distinction from some competitors. For example, Browserbase uses Exa to power its Search endpoint, meaning that layer is not proprietary. Firecrawl offers search, crawl, and an agent endpoint, but the agent endpoint has reliability issues on many tasks.<\/p>\nThe infrastructure argument is not only about avoiding vendor dependencies. When every layer of the stack is owned by the same team, the system can optimize for a single outcome: whether the task completed. When TinyFish\u2019s agent succeeds or fails using its own search and fetch, the company gets end-to-end signal at every step \u2014 what was searched, what was fetched, and exactly where failures occurred. Companies whose search or fetch layer runs on a third-party API do not have access to this signal.<\/p>\n
There is also a practical cost that teams integrating multiple providers encounter. Search finds a page the fetch layer cannot render. Fetch returns content the agent cannot parse. Browser sessions drop context between steps. The result is custom glue code, retry logic, fallback handlers, and validation layers \u2014 engineering work that adds up. A unified stack removes the component boundaries where these failures occur.<\/p>\n
The platform also maintains session consistency across steps: same IP, same fingerprint, same cookies throughout a workflow. Separate tools operating independently appear to a target site as multiple unrelated clients, which increases the likelihood of detection and session failure.<\/p>\n
Key Metrics<\/strong><\/h3>\n\n![\"\"](\"https:\/\/www.marktechpost.com\/wp-content\/uploads\/2026\/04\/image-30.png\")
<\/figure>\n<\/div>\n\n![\"\"](\"https:\/\/www.marktechpost.com\/wp-content\/uploads\/2026\/04\/image-31.png\")
<\/figure>\n<\/div>\n\n![\"\"](\"https:\/\/www.marktechpost.com\/wp-content\/uploads\/2026\/04\/image-29.png\")
<\/figure>\n<\/div>\nKey Takeaways<\/strong><\/h3>\n\n- TinyFish moves from a single web agent to a four-product platform \u2014 Web Agent, Web Search, Web Browser, and Web Fetch \u2014 all accessible under one API key and one credit system, eliminating the need to manage multiple providers.<\/li>\n
- The CLI + Agent Skill combination lets AI coding agents use the live web autonomously \u2014 install once and agents like Claude Code, Cursor, and Codex automatically know when and how to call each TinyFish endpoint, with no manual integration code.<\/li>\n
- CLI-based operations produce 87% fewer tokens per task than MCP, and write output directly to the filesystem instead of dumping it into the agent\u2019s context window \u2014 keeping context clean across multi-step workflows.<\/li>\n
- Every layer of the stack \u2014 Search, Fetch, Browser, and Agent \u2014 is built in-house, giving end-to-end signals when a task succeeds or fails, a data feedback loop that cannot be replicated by assembling third-party APIs.<\/li>\n
- TinyFish maintains a single session identity across an entire workflow \u2014 same IP, fingerprint, and cookies \u2014 whereas separate tools appear to target sites as multiple unrelated clients, increasing detection risk and failure rates.<\/li>\n<\/ul>\n
Getting Started<\/strong><\/h3>\nTinyFish offers 500 free steps with no credit card required at tinyfish.ai<\/a>. The open-source cookbook and Skill files are available at github.com\/tinyfish-io\/tinyfish-cookbook<\/a>, and CLI documentation is at <\/a>docs.tinyfish.ai\/cli<\/a>.<\/p>\n
\nNote:<\/strong> Thanks for the leadership at Tinyfish for supporting and providing details for this article.<\/em><\/p>\nThe post TinyFish AI Releases Full Web Infrastructure Platform for AI Agents: Search, Fetch, Browser, and Agent Under One API Key<\/a> appeared first on MarkTechPost<\/a>.<\/p>","protected":false},"excerpt":{"rendered":"AI agents struggle with tasks …<\/p>\n","protected":false},"author":1,"featured_media":728,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-727","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=727"}],"version-history":[{"count":0,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/media\/728"}],"wp:attachment":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=727"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=727"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=727"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
The Token Problem in Agent Pipelines<\/strong><\/h3>\nOne of the consistent performance problems in agent pipelines is context window pollution. When an AI agent uses a standard web fetch tool, it typically pulls the entire page \u2014 including thousands of tokens of navigation elements, ad code, and boilerplate markup \u2014 and puts all of it into the model\u2019s context window before reaching the actual content.<\/p>\n
TinyFish<\/a> Fetch addresses this by rendering the page in a full browser and returning only the clean text content as Markdown or JSON. The company\u2019s benchmarks show CLI-based operations using approximately 100 tokens per operation versus roughly 1,500 tokens when routing the same workflow over MCP \u2014 an 87% reduction per operation.<\/p>\nBeyond token count, there is an architectural difference worth understanding: MCP operations return output directly into the agent\u2019s context window. The TinyFish CLI writes output to the filesystem, and the agent reads only what it needs. This keeps the context window clean across multi-step tasks and enables composability through native Unix pipes and redirects \u2014 something that is not possible with sequential MCP round-trips.<\/p>\n
On complex multi-step tasks, TinyFish reports 2\u00d7 higher task completion rates using CLI + Skills compared to MCP-based execution.<\/p>\n
The CLI and Agent Skill System<\/strong><\/h3>\nTinyFish<\/a> is shipping two developer-facing components alongside the API endpoints.<\/p>\nThe CLI<\/strong> installs with a single command:<\/strong><\/p>\n\n\n\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nCopy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\nnpm install -g @tiny-fish\/cli<\/code><\/pre>\n<\/div>\n<\/div>\nThis gives terminal access to all four endpoints \u2014 Search, Fetch, Browser, and Agent \u2014 directly from the command line.<\/p>\n
The Agent Skill<\/strong> is a markdown instruction file (SKILL.md) that teaches AI coding agents \u2014 including Claude Code, Cursor, Codex, OpenClaw, and OpenCode \u2014 how to use the CLI. Install it with<\/strong>:<\/p>\n\n\n\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nCopy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\nnpx skills add https:\/\/github.com\/tinyfish-io\/skills --skill tinyfish<\/code><\/pre>\n<\/div>\n<\/div>\nOnce installed, the agent learns when and how to call each TinyFish endpoint without manual SDK integration or configuration. A developer can ask their coding agent to \u201cget competitor pricing from these five sites,\u201d and the agent autonomously recognizes the TinyFish skill, calls the appropriate CLI commands, and writes structured output to the filesystem \u2014 without the developer writing integration code.<\/p>\n
The company also notes that MCP remains supported. The positioning is that MCP is suited for discovery, while CLI + Skills is the recommended path for heavy-duty, multi-step web execution.<\/p>\n
Why a Unified Stack?<\/strong><\/h3>\nTinyFish<\/a> built Search, Fetch, Browser, and Agent entirely in-house. This is a meaningful distinction from some competitors. For example, Browserbase uses Exa to power its Search endpoint, meaning that layer is not proprietary. Firecrawl offers search, crawl, and an agent endpoint, but the agent endpoint has reliability issues on many tasks.<\/p>\nThe infrastructure argument is not only about avoiding vendor dependencies. When every layer of the stack is owned by the same team, the system can optimize for a single outcome: whether the task completed. When TinyFish\u2019s agent succeeds or fails using its own search and fetch, the company gets end-to-end signal at every step \u2014 what was searched, what was fetched, and exactly where failures occurred. Companies whose search or fetch layer runs on a third-party API do not have access to this signal.<\/p>\n
There is also a practical cost that teams integrating multiple providers encounter. Search finds a page the fetch layer cannot render. Fetch returns content the agent cannot parse. Browser sessions drop context between steps. The result is custom glue code, retry logic, fallback handlers, and validation layers \u2014 engineering work that adds up. A unified stack removes the component boundaries where these failures occur.<\/p>\n
The platform also maintains session consistency across steps: same IP, same fingerprint, same cookies throughout a workflow. Separate tools operating independently appear to a target site as multiple unrelated clients, which increases the likelihood of detection and session failure.<\/p>\n
Key Metrics<\/strong><\/h3>\n\n<\/figure>\n<\/div>\n\n<\/figure>\n<\/div>\n\n<\/figure>\n<\/div>\nKey Takeaways<\/strong><\/h3>\n\n- TinyFish moves from a single web agent to a four-product platform \u2014 Web Agent, Web Search, Web Browser, and Web Fetch \u2014 all accessible under one API key and one credit system, eliminating the need to manage multiple providers.<\/li>\n
- The CLI + Agent Skill combination lets AI coding agents use the live web autonomously \u2014 install once and agents like Claude Code, Cursor, and Codex automatically know when and how to call each TinyFish endpoint, with no manual integration code.<\/li>\n
- CLI-based operations produce 87% fewer tokens per task than MCP, and write output directly to the filesystem instead of dumping it into the agent\u2019s context window \u2014 keeping context clean across multi-step workflows.<\/li>\n
- Every layer of the stack \u2014 Search, Fetch, Browser, and Agent \u2014 is built in-house, giving end-to-end signals when a task succeeds or fails, a data feedback loop that cannot be replicated by assembling third-party APIs.<\/li>\n
- TinyFish maintains a single session identity across an entire workflow \u2014 same IP, fingerprint, and cookies \u2014 whereas separate tools appear to target sites as multiple unrelated clients, increasing detection risk and failure rates.<\/li>\n<\/ul>\n
Getting Started<\/strong><\/h3>\nTinyFish offers 500 free steps with no credit card required at tinyfish.ai<\/a>. The open-source cookbook and Skill files are available at github.com\/tinyfish-io\/tinyfish-cookbook<\/a>, and CLI documentation is at <\/a>docs.tinyfish.ai\/cli<\/a>.<\/p>\n
\nNote:<\/strong> Thanks for the leadership at Tinyfish for supporting and providing details for this article.<\/em><\/p>\nThe post TinyFish AI Releases Full Web Infrastructure Platform for AI Agents: Search, Fetch, Browser, and Agent Under One API Key<\/a> appeared first on MarkTechPost<\/a>.<\/p>","protected":false},"excerpt":{"rendered":"AI agents struggle with tasks …<\/p>\n","protected":false},"author":1,"featured_media":728,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-727","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=727"}],"version-history":[{"count":0,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/media\/728"}],"wp:attachment":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=727"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=727"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=727"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
Beyond token count, there is an architectural difference worth understanding: MCP operations return output directly into the agent\u2019s context window. The TinyFish CLI writes output to the filesystem, and the agent reads only what it needs. This keeps the context window clean across multi-step tasks and enables composability through native Unix pipes and redirects \u2014 something that is not possible with sequential MCP round-trips.<\/p>\n
On complex multi-step tasks, TinyFish reports 2\u00d7 higher task completion rates using CLI + Skills compared to MCP-based execution.<\/p>\n
The CLI and Agent Skill System<\/strong><\/h3>\nTinyFish<\/a> is shipping two developer-facing components alongside the API endpoints.<\/p>\nThe CLI<\/strong> installs with a single command:<\/strong><\/p>\n\n\n\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nCopy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\nnpm install -g @tiny-fish\/cli<\/code><\/pre>\n<\/div>\n<\/div>\nThis gives terminal access to all four endpoints \u2014 Search, Fetch, Browser, and Agent \u2014 directly from the command line.<\/p>\n
The Agent Skill<\/strong> is a markdown instruction file (SKILL.md) that teaches AI coding agents \u2014 including Claude Code, Cursor, Codex, OpenClaw, and OpenCode \u2014 how to use the CLI. Install it with<\/strong>:<\/p>\n\n\n\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\nCopy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\nnpx skills add https:\/\/github.com\/tinyfish-io\/skills --skill tinyfish<\/code><\/pre>\n<\/div>\n<\/div>\nOnce installed, the agent learns when and how to call each TinyFish endpoint without manual SDK integration or configuration. A developer can ask their coding agent to \u201cget competitor pricing from these five sites,\u201d and the agent autonomously recognizes the TinyFish skill, calls the appropriate CLI commands, and writes structured output to the filesystem \u2014 without the developer writing integration code.<\/p>\n
The company also notes that MCP remains supported. The positioning is that MCP is suited for discovery, while CLI + Skills is the recommended path for heavy-duty, multi-step web execution.<\/p>\n
Why a Unified Stack?<\/strong><\/h3>\nTinyFish<\/a> built Search, Fetch, Browser, and Agent entirely in-house. This is a meaningful distinction from some competitors. For example, Browserbase uses Exa to power its Search endpoint, meaning that layer is not proprietary. Firecrawl offers search, crawl, and an agent endpoint, but the agent endpoint has reliability issues on many tasks.<\/p>\nThe infrastructure argument is not only about avoiding vendor dependencies. When every layer of the stack is owned by the same team, the system can optimize for a single outcome: whether the task completed. When TinyFish\u2019s agent succeeds or fails using its own search and fetch, the company gets end-to-end signal at every step \u2014 what was searched, what was fetched, and exactly where failures occurred. Companies whose search or fetch layer runs on a third-party API do not have access to this signal.<\/p>\n
There is also a practical cost that teams integrating multiple providers encounter. Search finds a page the fetch layer cannot render. Fetch returns content the agent cannot parse. Browser sessions drop context between steps. The result is custom glue code, retry logic, fallback handlers, and validation layers \u2014 engineering work that adds up. A unified stack removes the component boundaries where these failures occur.<\/p>\n
The platform also maintains session consistency across steps: same IP, same fingerprint, same cookies throughout a workflow. Separate tools operating independently appear to a target site as multiple unrelated clients, which increases the likelihood of detection and session failure.<\/p>\n
Key Metrics<\/strong><\/h3>\n\n<\/figure>\n<\/div>\n\n<\/figure>\n<\/div>\n\n<\/figure>\n<\/div>\nKey Takeaways<\/strong><\/h3>\n\n- TinyFish moves from a single web agent to a four-product platform \u2014 Web Agent, Web Search, Web Browser, and Web Fetch \u2014 all accessible under one API key and one credit system, eliminating the need to manage multiple providers.<\/li>\n
- The CLI + Agent Skill combination lets AI coding agents use the live web autonomously \u2014 install once and agents like Claude Code, Cursor, and Codex automatically know when and how to call each TinyFish endpoint, with no manual integration code.<\/li>\n
- CLI-based operations produce 87% fewer tokens per task than MCP, and write output directly to the filesystem instead of dumping it into the agent\u2019s context window \u2014 keeping context clean across multi-step workflows.<\/li>\n
- Every layer of the stack \u2014 Search, Fetch, Browser, and Agent \u2014 is built in-house, giving end-to-end signals when a task succeeds or fails, a data feedback loop that cannot be replicated by assembling third-party APIs.<\/li>\n
- TinyFish maintains a single session identity across an entire workflow \u2014 same IP, fingerprint, and cookies \u2014 whereas separate tools appear to target sites as multiple unrelated clients, increasing detection risk and failure rates.<\/li>\n<\/ul>\n
Getting Started<\/strong><\/h3>\nTinyFish offers 500 free steps with no credit card required at tinyfish.ai<\/a>. The open-source cookbook and Skill files are available at github.com\/tinyfish-io\/tinyfish-cookbook<\/a>, and CLI documentation is at <\/a>docs.tinyfish.ai\/cli<\/a>.<\/p>\n
\nNote:<\/strong> Thanks for the leadership at Tinyfish for supporting and providing details for this article.<\/em><\/p>\nThe post TinyFish AI Releases Full Web Infrastructure Platform for AI Agents: Search, Fetch, Browser, and Agent Under One API Key<\/a> appeared first on MarkTechPost<\/a>.<\/p>","protected":false},"excerpt":{"rendered":"AI agents struggle with tasks …<\/p>\n","protected":false},"author":1,"featured_media":728,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-727","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=727"}],"version-history":[{"count":0,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/media\/728"}],"wp:attachment":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=727"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=727"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=727"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
The CLI<\/strong> installs with a single command:<\/strong><\/p>\n This gives terminal access to all four endpoints \u2014 Search, Fetch, Browser, and Agent \u2014 directly from the command line.<\/p>\n The Agent Skill<\/strong> is a markdown instruction file (SKILL.md) that teaches AI coding agents \u2014 including Claude Code, Cursor, Codex, OpenClaw, and OpenCode \u2014 how to use the CLI. Install it with<\/strong>:<\/p>\n Once installed, the agent learns when and how to call each TinyFish endpoint without manual SDK integration or configuration. A developer can ask their coding agent to \u201cget competitor pricing from these five sites,\u201d and the agent autonomously recognizes the TinyFish skill, calls the appropriate CLI commands, and writes structured output to the filesystem \u2014 without the developer writing integration code.<\/p>\n The company also notes that MCP remains supported. The positioning is that MCP is suited for discovery, while CLI + Skills is the recommended path for heavy-duty, multi-step web execution.<\/p>\n TinyFish<\/a> built Search, Fetch, Browser, and Agent entirely in-house. This is a meaningful distinction from some competitors. For example, Browserbase uses Exa to power its Search endpoint, meaning that layer is not proprietary. Firecrawl offers search, crawl, and an agent endpoint, but the agent endpoint has reliability issues on many tasks.<\/p>\n The infrastructure argument is not only about avoiding vendor dependencies. When every layer of the stack is owned by the same team, the system can optimize for a single outcome: whether the task completed. When TinyFish\u2019s agent succeeds or fails using its own search and fetch, the company gets end-to-end signal at every step \u2014 what was searched, what was fetched, and exactly where failures occurred. Companies whose search or fetch layer runs on a third-party API do not have access to this signal.<\/p>\n There is also a practical cost that teams integrating multiple providers encounter. Search finds a page the fetch layer cannot render. Fetch returns content the agent cannot parse. Browser sessions drop context between steps. The result is custom glue code, retry logic, fallback handlers, and validation layers \u2014 engineering work that adds up. A unified stack removes the component boundaries where these failures occur.<\/p>\n The platform also maintains session consistency across steps: same IP, same fingerprint, same cookies throughout a workflow. Separate tools operating independently appear to a target site as multiple unrelated clients, which increases the likelihood of detection and session failure.<\/p>\n TinyFish offers 500 free steps with no credit card required at tinyfish.ai<\/a>. The open-source cookbook and Skill files are available at github.com\/tinyfish-io\/tinyfish-cookbook<\/a>, and CLI documentation is at <\/a>docs.tinyfish.ai\/cli<\/a>.<\/p>\n Note:<\/strong> Thanks for the leadership at Tinyfish for supporting and providing details for this article.<\/em><\/p>\n The post TinyFish AI Releases Full Web Infrastructure Platform for AI Agents: Search, Fetch, Browser, and Agent Under One API Key<\/a> appeared first on MarkTechPost<\/a>.<\/p>","protected":false},"excerpt":{"rendered":" AI agents struggle with tasks …<\/p>\n","protected":false},"author":1,"featured_media":728,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-727","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=727"}],"version-history":[{"count":0,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/727\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/media\/728"}],"wp:attachment":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=727"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=727"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=727"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}npm install -g @tiny-fish\/cli<\/code><\/pre>\n<\/div>\n<\/div>\nnpx skills add https:\/\/github.com\/tinyfish-io\/skills --skill tinyfish<\/code><\/pre>\n<\/div>\n<\/div>\nWhy a Unified Stack?<\/strong><\/h3>\n
Key Metrics<\/strong><\/h3>\n
<\/figure>\n<\/div>\n<\/figure>\n<\/div>\n<\/figure>\n<\/div>\nKey Takeaways<\/strong><\/h3>\n
\n
Getting Started<\/strong><\/h3>\n
\n