Alma by Olivares.AI

Alma is the AI assistant that remembers every conversation, every preference, every detail you share — across every session, every project, every modality.

One subscription. One weekly budget. One workspace for chat, image, video, music, documents, calendar and search.

https://youtu.be/SmLvd4URXb4

What Alma gives you:

• Persistent memory across all your conversations — facts, preferences, decisions, episodes

• Image, video and music generation under a single weekly budget

• Calendar, documents and web search built into the chat

• Writing tools, templates, skills, folders — everything you create stays organized

• BYOK and MCP server (Claude Desktop / Cursor) on the Max plan

Plans:

Starter — $14/month

Pro — $29/month

Max — $99/month

Try it: https://alma.olivares.ai

More info: https://olivares.ai

New Pricing: Starter, Pro, Max — One Weekly Budget, Every Feature

Simplified from seven plans to three: Starter, Pro and Max. Every paid plan includes a single weekly USD budget that covers chat, voice, images, video and music from the same pool. Top up any time. Memory, images, voice, video, studio-grade music, code, writing tools, web search and export are all included on every paid plan — nothing to bolt on. Annual billing available with two months off. Your data is preserved indefinitely — cancel any time and resume exactly where you left off.

https://youtu.be/f1hVzsqP2TM

Three plans, one budget

• Starter — persistent memory with Haiku and Sonnet, all tools, web + mobile apps

• Pro — adds the top reasoning tier (Claude Opus), video, studio-grade music and a larger weekly budget

• Max — everything in Pro plus the full developer surface: API keys, VSCode extension, MCP server, JavaScript SDK and BYOK support

• Annual billing on every plan: pay ten months, get twelve — two months off for the year

• A single USD pool per week covers chat, voice, images, video and music. Resets every Monday 00:00 UTC

Top-ups when you need more

• Top up when the weekly pool runs low and you opt in — nothing auto-charges

• Top-up balance expires 12 months after purchase and is consumed in FIFO order

• Usage breakdown by category (chat, image, voice, video, music) visible in Settings → Billing

Your data stays, no matter what

• Every memory, conversation and file is preserved indefinitely — no purge on cancellation

• Read-only access to everything you created, even without an active subscription

• Resubscribe any time and resume exactly where you left off

• Full export at any time: JSON, .alma, PDF, DOCX, XLSX, HTML and Markdown

Cleaner product surface

• Consolidated seven legacy plans down to three — less decision fatigue, clearer trade-offs

• Music and video integrated into the same weekly budget — no separate add-on subscriptions

• Pricing page rewritten with a compact plan grid, full feature comparison and top-up preview

• Settings → Billing shows live weekly budget, category usage and one-click top-ups

Developer surface on Max

• Programmatic API with X-API-Key authentication

• VSCode extension and MCP server (Claude Desktop, Cursor, Windsurf)

• JavaScript SDK published to npm (@olivaresai/alma-sdk)

• Bring Your Own Key (Anthropic, Replicate, Leonardo AI) — your BYOK calls do not consume your Alma budget

Alma is an AI workspace built around persistent memory. Unlike standard AI chatbots that reset every session, Alma learns your preferences, decisions, and working patterns — and applies that context across every tool in the platform.

The Creator plan ($89/month) bundles:

• AI Chat powered by Anthropic Claude — with persistent memory that accumulates over time (50,000 memories, knowledge graph, semantic search)

• Video Studio — Runway Gen-4 Turbo + Gen-4.5 with 300 credits/month, 12 camera movements, 7 aspect ratios

• Music Studio — ElevenLabs Premium (up to 10 min/track with vocals and custom lyrics) + MiniMax for quick ideas

• Image Studio — Leonardo AI + Flux Pro, 10 style presets, gallery, public sharing

• 12 tools total — including web search, code workspace, deep research, document generation (DOCX/XLSX/PPTX/PDF), BYOK, REST API, MCP server, VSCode extension

Product demo (3:30): https://youtu.be/sQzsiBBiAOg

Why one platform instead of separate tools?

The real problem with using ChatGPT + Runway + ElevenLabs + Midjourney separately isn't cost — it's context fragmentation. You describe your brand in one tool, then repeat yourself in the next. None of them share memory.

Alma's persistent memory connects every creative tool. When you describe your project in chat, generate images from that conversation, turn images into video, and compose a soundtrack — the AI maintains context across the entire workflow.

Every other AI tool starts from zero every session. Alma compounds context over time.

The differentiator: Soul Engine

Beyond memory, Alma includes a Soul Engine — a structured identity system for your AI with configurable blocks: identity, personality, expertise, rules, worldview, anti-patterns, communication modes. This isn't a system prompt that gets ignored after 10 messages. It's a versioned, persistent personality layer that shapes every response.

Try it: https://alma.olivares.ai

Creator plan: https://olivares.ai/creator/

Docs: https://olivares.ai/docs

https://youtu.be/sQzsiBBiAOg

I've been building Alma — an AI assistant that remembers everything about you across conversations.

Just published a full product demo showing all the features:

https://youtu.be/XeIxJLzcRLc

https://olivares.ai/demo

The highlight: I give Alma a complex planning prompt and it automatically searches through my past memories and conversations to generate a complete professional document — without me repeating any context.

What's inside:

- Chat with 3-layer persistent memory

- Soul Engine (configure your AI's personality)

- Video Studio (Runway Gen-4)

- Image Studio (Flux Pro)

- Writing Tools (7 AI transformations)

- Web Search with Deep Research

- Documents, Ideas, Dashboard

Stack: Cloudflare edge (Workers, D1, Vectorize, Durable Objects), React 19, Anthropic Claude. 2,964 tests. Solo dev.

Free tier at alma.olivares.ai — would love feedback from this community.

Building Alma as a solo dev. This week I shipped new update with 10+ new features.

The big ones:

1. Code — AI workspace for code. Clone repos, explore files, run 8 skills (explain, refactor, review...). Like a mini Claude Code in the browser.

2. Video Studio — Runway Gen-4/4.5 integration. Plan scenes with AI, stitch clips, generate YouTube metadata. Credit-based billing.

3. Global Search — Search across all conversations. Finally find "that thing I said last week."

4. Conversation Branching — Fork from any message. Explore two approaches without losing either.

Revenue update:

• 2 Pro subscriber ($19/mo)

• Hundreds of visits from TAAFT

• Dozens of free signups

Lesson learned: Ship the core feature first (memory), then expand. Users came for the memory system, stayed for Image Studio and Chat. Now Code and Video give them reasons to upgrade.

I've been building Alma for months. It's a persistent memory and identity engine for AI — web app, MCP server, VSCode extension, REST API. I use it daily and I'm at the point where going back to vanilla Claude feels broken.

Here's what Alma actually does.

When you chat, a background processor fires every few messages. It extracts structured facts from the conversation — not a raw transcript dump, but discrete memories. Each one gets a category (preference, decision, project, fact), an importance score (0 to 1), and a confidence level (1.0 if you stated it directly, 0.7 if the AI inferred it). Duplicates merge automatically. Important things rise, irrelevant things fade.

That's the first layer. The second is episodes — compressed conversation summaries. Not "you talked for 45 minutes" but "debugged auth middleware, root cause was missing await, decided to add retry logic." Next time auth comes up, the AI already has that context.

The third is procedures — behavioral patterns the AI observes over time. "Checks error handling first in code reviews." "Prefers minimal comments in code." You never write these. They emerge from how you work.

All three layers feed into a 12-block identity system I call the Soul Engine. It separates who the AI is (stable identity) from how it communicates (style that evolves) from what it knows about you (context that auto-updates). The single highest-impact piece: the anti-patterns block, where you define what the AI must never do. Five rules there improved my output more than pages of positive instructions.

The tech: Cloudflare Workers, D1, Vectorize, R2, Durable Objects, Workers AI for embeddings. Hono framework. React 19 frontend. 2,622 tests across 149 files, 69 database migrations. Claude-only (Haiku free, Sonnet + Opus paid). Hybrid search combining keyword matching with vector similarity.

The honest part: I'm a solo developer. I launched on Product Hunt and TAAFT two days ago. I have a handful of paying subscribers and I'm figuring out distribution. The product works — the challenge is getting it in front of the right people.

If you use AI daily for work, the value kicks in around week two when the accumulated context starts compounding. The first session feels normal. The twentieth feels like a different product.

alma.olivares.ai — free tier, no credit card. Feedback on the product or the approach is genuinely welcome.