You use Claude, GPT, and DeepSeek. Every session starts from zero.
DARA is a file they all share — write once, read everywhere.
Open source. Local. No cloud required.
You open a new chat with Claude. You paste your project context.
You switch to GPT. You paste it again. DeepSeek? Paste.
Week two: you have 4 versions of the same file and you're not sure which is current.
Your AIs are brilliant. But every conversation is a first date.
A shared memory file that any AI can read and write. A compiler keeps it clean.
Any AI writes.
Any platform.
Follow the rules.
10-step pipeline.
Validates. Dedupes.
Auto-fixes. Locks.
Any AI reads.
One file. Current.
SHA256 verified.
Write anywhere. Compile once. Read everywhere.
No terminals. No file management. Just talk to your AI.
→ Consult DARA
Open any AI and say:
"Check DARA for context on project X."
Seconds later, the AI knows everything.
→ Write to DARA
When something new happens:
"Update DARA with what we just decided."
The AI writes it. Next session, it's there.
If you checked 2 or more, EIDARA will save you time starting today.
4 AI models. 52 tests each. No human intervention. No cherry-picking.
No other memory system publishes cross-model test results.
We do because it has to work with your AI — not just ours.
A 15-rule constitution governs every write. Any AI that reads the rules can participate. No training, no permissions — the rules ARE the access control.
A 10-step pipeline validates, deduplicates, auto-fixes, and checksums your memory. No other system compiles — they just accumulate.
Errors are expected. The next AI that spots one fixes it automatically. Consensus flags protect against unilateral changes to important data.
Markdown files + Python standard library. No database, no Docker, no API keys, no cloud account. Runs anywhere Python 3.10+ exists.
Before DARA worked, it failed three times — as task experts, role permissions, and folder hierarchies.
A mail agent doesn't understand your business. A domain expert does. Specialized agents fail because context is cross-cutting.
Only authorized AIs can write. But any AI might have the right info. Gatekeeping creates bottlenecks, not quality.
Organizing memory for humans who browse folders. But humans never open the folder — the AI does. Organize for the reader.
"If the human never opens the folder, who are you organizing for?"
That question changed everything.Each topic lives in its own file — called a neuron. One project = one file. One tool = one file. Plain markdown that any AI understands.
Drop a neuron in shared Drive. Your colleague opens it with their AI — Claude, GPT, DeepSeek, doesn't matter. No install needed. It's just a text file.
Agents work the same way. A file contains the protocol. Any AI that reads it becomes that agent. The intelligence lives in the file, not in who runs it.
# agent-librarian.md
summary: Vault maintenance agent
## Protocol
1. Scan INBOX for pending items
2. Route each to correct neuron
3. Flag ambiguous entries
4. Log actions in changelog
$ git clone https://github.com/jrotllant/eidara.git
$ cd eidara
$ python compile.py
Or give the INSTALL file to any AI. It guides you through 11 steps — personalization, git setup, and watcher configuration. You don't type a single terminal command.
EIDARA is free. MIT licensed. No telemetry. No cloud. Your data never leaves your machine.
Built by Javier Rotllant — pragmatic founder, Bain background, not an engineer. Someone with a problem and the stubbornness to solve it properly.