Independent review. Unpaid; no partnership with Consensus. Based on public documentation and research. Trademarks identify their owners’ products. Disclosure.
Searches scholarly literature and summarizes findings with source papers. Its MCP server connects search to ChatGPT, Claude and supported clients; plan limits control result depth and shared API/MCP usage. Check the papers and whether they are peer reviewed.
The Ion perspective
The five-point review
Judgment, not a score. Documentation checked 2026-10-18.
Value
Searches scholarly literature and summarizes findings with source papers. Its MCP server connects search to ChatGPT, Claude and supported clients; plan limits control result depth and shared API/MCP usage. Check the papers and whether they are peer reviewed.
Differentiation
Documented fit: Fact-checking claims against published research. Uniqueness is not established.
Replaceability and switching
Compare Elicit, Perplexity. Not drop-in replacements. Export completeness and migration effort remain unverified.
Ease and technical aptitude
Consider the fit limits: General web or market research. Setup effort has not been scored.
Interoperability and portability
See documented MCP, API and import routes below. Import does not prove export; API access does not prove a complete migration path. Check action scope, plans and portability.
Scholarly search inside AI assistants over MCP source ↗
Run deeper literature reviews (3 a month on Free, 15 on Pro, 200 on Deep) source ↗
Compare findings across uploaded research PDFs in a Collection source ↗
In depth
Plans
Pro costs US$20/month or US$144/year (US$12/month equivalent); Deep costs US$65/month or US$540/year (US$45/month equivalent). Free/Pro/Deep include 3/15/200 Deep reviews monthly and 30/500/2,000 shared API/MCP calls. Additional usage on eligible paid plans must be enabled and costs US$0.05 per call, subject to plan maximums. Checked 2026-10-18. source ↗
MCP access
Unauthenticated search returns a limited result set. Signed-in/API-key access follows the Consensus plan, independently of the AI assistant subscription. API and MCP share one monthly pool; searches count one call per 100 papers returned, rounded up, while research Threads may consume multiple calls. source ↗
Uploaded documents
Upload up to 20 PDFs at a time and ask questions across documents in a Collection. Text-based PDFs work best; scanned/image PDFs may fail, and the file-size limit is 50 MB. Uploads are private unless a Collection containing them is shared. source ↗
How to use it with AI
How Consensus connects to each AI platform: native connector or MCP, API through code the agent writes, browser control, or file upload.
Not documentedOpenAI Codex: Edits files and runs commands locally; cloud tasks run in isolated workspaces Consensus publishes an API, but this pairing still needs network, authentication and execution-scope checks.Consensus API ↗OpenAI Codex docs ↗
InferredOnly in the ChatGPT desktop app; not in the Codex CLI or IDE extension. Consensus is a web app; the pairing itself is not documented.OpenAI Codex browser docs ↗
Not documentedVerify which side exports and which imports, supported formats and permissions. Consensus's recorded upload/import status is yes; OpenAI Codex's file status is yes. These flags do not establish a file-transfer route.Consensus file docs ↗OpenAI Codex file docs ↗
InferredThe /browser guide documents App and CLI support through Chrome DevTools MCP; the overview comparison conflicts on CLI availability. IDE browser features are separate. Consensus is a web app; the pairing itself is not documented.Google Antigravity browser docs ↗
Not documentedVerify which side exports and which imports, supported formats and permissions. Consensus's recorded upload/import status is yes; Google Antigravity's file status is yes. These flags do not establish a file-transfer route.Consensus file docs ↗Google Antigravity file docs ↗
Confirmed is reserved for evidence of the specific connection route. The current named integrations include plugins, skills and CLIs, so a named platform alone does not confirm MCP or a native connector. Inferred means each side documents a relevant capability, but no document describes the pairing. Code execution still needs network and credential access; file handling still needs a compatible format and transfer direction. Not documented means the records do not establish the route. IonBin has not tested these pairings. Recorded Oct 18, 2026.
Automation hubs to assess for a custom API connection: Gumloop, Make, n8n, Zapier. Verify credentials, supported requests, limits and error handling; no configured integration is implied.
AI research assistant for finding papers, extracting data into tables and summarizing evidence across more than 138 million papers. The Elicit API and MCP server (Pro, Scale and Enterprise plans) bring search, reports and systematic reviews into Claude and other MCP clients.
Research assistant with source citations, paid model options and a separate Computer product for tasks such as running code. Its official MCP server lets other assistants call Perplexity search and research; that API usage is billed separately from consumer subscriptions.