Research project
Frida
This site describes the public state of FridaDev and signals some ongoing work around Frida.
The project
Current state of the project
Frida is a research project on the architecture of dialogue systems with memory, documents, validation mechanisms, and revision procedures.
The project treats the user prompt as a text and the pipeline as the technical place where that text meets memory traces, documents, and revision operations. It does not assume that a model understands in any strong sense; it works on the conditions under which a dialogue can be maintained, reread, corrected, or suspended.
At present, FridaDev is the only public application repository. It is a Python/Flask web runtime with browser chat, memory, a hermeneutic branch, a public streaming contract, and administration or observability surfaces.
Frida_V4 is ongoing work and is not public at this time. It is a native macOS runtime in Swift, organised by modules. The site may refer to some of its lines of work, but the repository itself is not currently publicly available.
This site, frida-ai, has a documentary function. It is used to describe the current state of the work, gather public documents, and distinguish what is available from what is not.
The problems treated here concern, among other things, dialogue continuity, memory organisation, the handling of injected documents, the revision of inferences, the status of hermeneutic or dialogic signals, forms of mutable system identity, and the conditions under which a response can be retained.
Research areas
01
FridaDev
Public repository. Python/Flask web runtime, memory, hermeneutic branch, streaming, administration surfaces, and operational documentation.
02
Frida_V4
Non-public ongoing work. Native macOS runtime in Swift, organised by modules, with memory, documents, voice, TTS, and local tooling.
03
Research problems
Memory, summaries, documents, arbitration, dialogic regime, persistence, revision, and the status of evidence in the treatment of a dialogue turn.
04
Public documents
Audits, cartographies, streaming contracts, operational documentation, and a working paper on the M6 module.
Architecture — overview
What the Architecture page describes
The Architecture page describes the current state of the project by distinguishing three levels: the public site, FridaDev as a public application repository, and Frida_V4 as non-public ongoing work. It does not present a single system available as a whole.
It also specifies the status of the modules or documents it mentions, in order to avoid confusion between what belongs to a public repository, an isolated public document, or work that remains closed.
Availability
What the site makes available
The site does not replace the repositories. It gives access to public documentation, to some working documents made public, and to a concise description of the current state of the project.
Status of the elements mentioned
The following statuses are distinguished on the site: