Skip to content

Confer vs Maple AI

Price, platforms and features side by side. The table says where each price was read and when.

Confer

End-to-end encrypted AI assistant from the creator of Signal

Free plan · from $34.99

Maple AI

End-to-end encrypted AI chat that runs open models in enclaves

Free plan · from $20/mo

Where they differ

  • Confer starts at $34.99 (Confer Membership, US App Store; billing period not shown). Maple AI starts at $20 per month (Pro).
  • Only Maple AI lists macOS, Windows, Linux and Android apps.
  • Maple AI is open source, according to its own site.
In commonFree planWebiOSEnd-to-end encryptedAI featuresChat assistantAI search

At a glance

ConferMaple AI
PricingFree planFree plan
Starts at$34.99 (Confer Membership, US App Store; billing period not shown)$20 per month (Pro)
Free planFree to start with no credit card: up to 5 chats and a daily message limit. Attachments, connectors and folder custom instructions are Member features.End-to-end encrypted chat with access to basic AI features. Web search is not included, and the pricing page does not state a usage allowance.
Free trialNone listedNone listed
Web
macOS
Windows
Linux
iOS
Android
Made byConfer Enterprises, IncMaple Privacy Labs Inc.
Price read fromapps.apple.comtrymaple.ai
CheckedOct 6, 2026Oct 9, 2026

Who each one suits

Confer

  • Discussing health, money or work without leaving a readable server-side record
  • People who trust Signal's approach and want the same for AI chat
  • Gmail, Google Drive and Google Calendar users who want private connectors

Maple AI

  • People who want cloud-hosted models without exposing their chats
  • Professionals working with confidential client or company material
  • Developers routing coding tools through an encrypted API

Worth knowing

Confer

  • The membership is listed at $34.99 as an in-app purchase on the US App Store with no billing period shown, and the website shows no price before sign-in.
  • The free plan is capped at 5 chats and a daily message limit; attachments, connectors and folder custom instructions need a membership.
  • It is available on the web and iPhone only. The App Store listing says the app is only for iPhone and requires iOS 26 or later.
  • Confer runs on an open-weight model that it selects; its site does not list models from OpenAI, Anthropic or Google.

Maple AI

  • Only open-weight models are offered; proprietary models from OpenAI, Anthropic or Google are not part of the lineup.
  • The Free plan leaves out web search, and uploads, voice recording and API access start at Pro for $20 per month.
  • API usage is charged per million tokens at each model's rate, separately from the subscription.
  • Maple Agent, the assistant with long-term memory, is in private beta behind a waitlist.

Key features

Confer

End-to-end encrypted history
Chat history, folders, memories and attachments are encrypted with keys derived from a passkey, so the server stores only ciphertext. Recovery keys can unlock the data on your other devices.
Private inference
Prompts travel over an encrypted channel into a confidential VM. The client checks the hardware attestation against a public transparency log, and the code for that environment is published.
Web search, images and file work
Answers can draw on web search with sources and on attached photos and files, and Confer can create or edit documents, spreadsheets, presentations and images or run code.
Memory you control
Persistent memory carries context from one conversation to the next. You can review, edit and delete what it remembers, and import conversations from ChatGPT.
Folders with files and instructions
Folders group conversations, carry their own custom instructions and hold PDFs, Word documents, spreadsheets and other files that every chat in the folder can reference.
Client-side connectors
Connectors for Gmail, Google Drive and Google Calendar run on your device: access tokens are stored encrypted and the results do not pass through Confer's servers in readable form.

Maple AI

End-to-end encrypted chats
Messages are encrypted on the device with a per-account key and decrypted only inside a hardware-isolated enclave for inference. Maple says it cannot read them and does not train on them.
Open models inside enclaves
Open-weight models such as Kimi, GLM, DeepSeek, Gemma, Llama and OpenAI's GPT-OSS run on GPUs in trusted execution environments, so prompts are not passed to the labs that made them.
Public code and attestation
Client and server code are published on GitHub under the MIT license. Reproducible builds and a signed enclave attestation let anyone check what is running.
Private web search
A search agent queries the web anonymously through Brave's search API and brings current information into the chat. Web search is left out of the Free plan.
Projects, documents and voice
Chats can be grouped into projects and pinned. Pro adds image and document upload in the apps, plus voice recording, inside the same encrypted workspace.
Maple Proxy API
An OpenAI-compatible endpoint that verifies the enclave and encrypts requests from tools such as OpenCode and Goose. API usage is billed per million tokens, apart from the plan price.

Plans

Confer

  • Free$0

    Up to 5 chats at a time and a daily message limit. Attachments, connectors and custom instructions for folders are reserved for Members.

  • Confer Membership$34.99

    in-app purchase on the US App Store; period not shown

    Removes the cap on chats and adds attachments, connectors for Gmail, Google Calendar and Google Drive, and custom instructions for folders.

Maple AI

  • Free$0

    End-to-end encrypted chat with basic AI features. Web search is not included.

  • Pro$20

    per month

    All models including Kimi, GLM, DeepSeek and GPT-OSS, image and document upload in the apps, voice recording and API access.

  • Team$30

    per user / month

    Everything in Pro with 2x more usage per member, chat credits pooled across the team and unified billing.

  • Max$100

    per month

    Everything in Pro with 20x more usage, priority support and early access to features and models.

More comparisons