FNEF working landscape
A practical picture of learner-facing apps and the wider systems that make them possible: dictionaries, readers, games, flashcards, keyboards, structured language-content and media workflows, collection, consent and preservation tools, collaboration services, APIs, AI and speech technology, data-governance principles and frameworks, research programs, and language-support and policy organizations. It is grounded in the Canadian First Nations language revitalization context while drawing on international and other-country projects to broaden the comparison, reveal different approaches, and surface ideas that may be useful in Canadian work. Prepared for communities, language teams, leaders, educators, program staff, technical partners, and language-support organizations.
How to read this page
This page is an environmental scan of what is available across learner-facing software and the wider language-revitalization technology, governance, and support ecosystem. It is grounded in the Canadian First Nations language revitalization context, including community authority, ownership of language content, data sovereignty, permissions, local staff capacity, connectivity, and long-term sustainability. It is not a ranking, a procurement recommendation, or a claim that one product, principle, framework, model, workflow, institution, or research program will fit every Nation or language organization.
“Low-resource” is a technical description of limited digital data, tools, or infrastructure. It is not a description of a language’s value, complexity, knowledge, or the strength of its community.
The entries are intentionally different. Some are complete applications; some generate standalone apps; some are desktop production tools, collaboration services, hosted platforms, APIs, AI model families, datasets, data-governance principles or frameworks, research partnerships, or language-support and policy bodies; and some are language-specific projects included because they demonstrate useful features. They should not be read as interchangeable products.
International and other-country examples are included deliberately. They widen the field of view by showing different governance models, public-investment strategies, technical architectures, teaching approaches, speech and AI methods, and ways of organizing long-term language work. They can offer transferable ideas and useful cautions for Canadian First Nations programs, but they are not assumed to transfer directly: every approach must still be interpreted through the language, culture, laws, protocols, priorities, and capacity of the Nation or organization considering it.
The central question is fit with the language program. Existing language content, workflow, staff capacity, governance, permissions, connectivity, teaching approach and long-term maintenance often matter more than the size of a feature list.
The directory can be used independently to build a shortlist. The FNEF workflow below is one concrete example of how people, tools, and governed sources can be connected into an end-to-end process.
FNEF context
The workflow begins with the community’s language priorities and the relationships through which language knowledge is shared, reviewed, taught, and protected. Language teams work with Elders, Language Keepers, and fluent speakers; FNEF linguists help organize and analyse language content; and software developers connect governed source material to the tools a program chooses to create.
Set priorities, exercise authority, define protocols and permissions, direct the work, and decide what may be shared.
Provide and review language, meaning, pronunciation, context, teaching knowledge, and cultural guidance.
Work alongside community teams to document, analyse, structure, and review language content, while training community language workers to use and sustain FLEx, ELAN and other workflow tools.
Build and maintain integrations and learner-facing tools around the governed language source and program goals.
FLEx serves as the single structured source of truth for reviewed lexical, grammatical, and text content in this workflow. It does not have to be the sole repository for every project asset: original recordings, preservation masters, permissions and consent records, governance decisions, and culturally sensitive materials remain in the community-controlled systems appropriate to them.
FNEF did not find a single off-the-shelf product that matched this FLEx-centred workflow and the goal of moving reviewed, structured language content into dictionaries, reading, games, lessons, archives, and custom applications. LanguageCloud was therefore co-designed with the communities it serves to fill that integration gap, not to replace every other tool. It includes a backend and APIs used for custom applications and integrations.
FNEF linguists train community language workers to use FLEx, ELAN and other workflow tools and to carry the structured documentation, review, media, collaboration, and publishing process forward. The goal is not permanent reliance on FNEF or another contractor. As local experience and technical confidence grow, community language workers can increasingly maintain, expand, and reuse their language resources with less external support.
This strengthens the language program by placing durable workflow skills in the hands of the people best situated to continue the work. Over time, the community retains not only its language data and resources, but also more of the practical capacity needed to steward them for future generations.
The workflow is deliberately modular. Tools can be replaced or supplemented as community requirements change, provided that the authoritative language content, media, identifiers, permissions, and structured exports remain under community control. AI and speech technology should reduce workload, increase access, and create safe opportunities to practise; it should not displace fluent speakers, teachers, families, cultural authority, or relationships. The intended direction is back into human language use.
The durable outcome is not only an app. It is a governed, reusable language-content system—and the relationships, skills, source materials, and community-held capacity required to maintain it. FNEF’s role is to connect people and specialist work, provide training, and help reduce long-term reliance on outside contractors, including FNEF itself—not to become the source of language authority. The same questions about fit, control, source materials, maintenance, and exit recur across the wider landscape.
Affiliation disclosure: FNEF publishes this scan and maintains LanguageCloud. LanguageCloud should be assessed by the same fit, governance, evidence, sustainability, and exit criteria as every other entry.
The strongest tools support people in speaking, listening, reading, teaching, and creating together. Digital engagement is most valuable when it encourages everyday use across ages, abilities, homes, schools, workplaces, lands, and community life.
Dictionaries, readers, games, keyboards, archives, annotation, collaboration, ASR, TTS, translation, and pronunciation are usually separate layers. Exact fit often comes from integration and governance rather than one purchase.
Projects from Aotearoa New Zealand, Hawaiʻi, Alaska, Sápmi, Australia, Africa, Asia, Latin America, and elsewhere reveal approaches that may not yet be visible in Canada. Their value is comparative: they can inform governance, funding, curriculum, software architecture, speech technology, and community participation without being treated as ready-made templates for a Canadian First Nations context.
FLEx, ELAN, Audacity, LexBox, Keyman, review procedures, naming conventions, permissions, and backups may be invisible to learners, but they determine whether content can be updated, reused, protected, and sustained. Long-term sustainability also depends on ensuring community language workers have the training, time, and authority to continue using these tools and processes.
A language technology project is sustainable only when community language workers have the skills, time, authority, and ongoing support needed to continue the work. Hands-on training, mentoring, documentation, paid learning time, and succession planning should be treated as essential infrastructure—not optional additions at the end of implementation. The goal is to build community-held capacity to operate, maintain, and adapt FLEx, ELAN and other workflow tools, progressively reducing reliance on outside contractors while keeping the knowledge and practical skills close to the language, its speakers, and future generations.
DAB, RAB, Bloom, AlphaTiles, and similar tools are strong for offline or bounded outputs. Web systems support collaborative updates, APIs, and shared services, but connectivity, hosting, security, and long-term operations matter.
Anki, H5P, Quizlet, Mnemosyne, and other practice systems can support daily habits, reusable activities, and spaced review. They do not replace curriculum, fluent-speaker relationships, communicative practice, or a maintained source of reviewed language content.
There are many dictionary, story, and vocabulary tools. Fewer systems connect structured lexical and grammatical content to sustained practice, teacher planning, assessment, and real-world communicative use.
Organizations such as FPCC, FNCCEC, SILR, OCIL, and the NWT Indigenous Languages and Education Secretariat influence research, policy, funding, training, rights, curriculum, convening, and institutional capacity. They belong in the landscape even though they are not software products.
Principles and frameworks such as OCAP® and Local Contexts help language programs define authority, responsibilities, access, possession, provenance, permissions, and culturally appropriate use. They belong in the landscape because these decisions must inform system design, agreements, metadata, and operations—not be added after deployment.
ASR, TTS, machine translation, LLMs, speech-text alignment, and pronunciation assessment solve different problems and require different data, evaluation, and safeguards.
A model that lists a language—or performs well on a global benchmark—may still fail on local dialects, spelling systems, age groups, code-switching, recording conditions, and culturally important meanings.
For polysynthetic and morphologically complex languages, FLEx structures, finite-state models, paradigms, and reviewed analyses can strengthen search, conjugation, learning activities, translation, and speech tools.
Keyboards, fonts, normalization, corpora, transcription, permissions, speaker metadata, audio standards, and stable identifiers are prerequisites. Model quality cannot compensate for weak or poorly governed foundations.
Governance must cover recordings, transcripts, training sets, prompts, embeddings, synthetic text, fine-tuned weights, generated voices, API logs, withdrawal, and what happens when a partnership ends.
Learner-facing mobile and web applications, PWAs, app builders, dictionaries, readers, games, flashcards, keyboards, hosted services, APIs, reusable frameworks, AI and speech systems, model families, datasets, collection, consent and preservation tools, desktop language-data and media workflows, collaboration infrastructure, data-governance principles and frameworks, research programs, and language-support or policy bodies that materially shape implementation.
The landscape is grounded in Canadian First Nations language revitalization work, but it intentionally includes Inuit, Métis, Indigenous and minoritized-language initiatives from Canada and internationally. These comparisons help surface possibilities, risks, governance approaches, and investment models that may inform Canadian work while respecting that local authority and context remain decisive.
General field collection, office software, learning-management systems, and desktop linguistic tools remain outside scope unless they directly support the FNEF workflow, publish an end-user experience, or materially enable a reusable language-content, speech, text, model, or application pipeline. General-purpose AI appears only where it has clear low-resource-language relevance.
Entries draw on public product pages, official documentation, source repositories, research and government program pages, governance-framework resources, and clearly labelled FNEF or user firsthand notes. A deployable tool, a governance principle or framework, a workflow component, an institution, an adaptable model, a research prototype, and a dataset are treated as different forms—not as equivalent products.
Costs, maintenance, language coverage, model accuracy, app-store availability, security, hosting, licences, institutional programs, and contract terms can change. AI claims must be tested with fluent speakers and intended users. Entries containing FNEF or user-provided observations are labelled as firsthand and should be independently reconfirmed where appropriate.
Filterable directory
Use the filters to identify candidates, then expand their details to assess fit, community control, implementation effort, evidence, and the path for exporting work or leaving a service. Filter by geography, entry type, function, AI or speech capability, delivery, offline support, language or governance-implementation pathway, source availability, implementation demands, or the tools used in the FNEF workflow.
See how community-led work moves through a governed FLEx source, media preparation, collaboration, and learner-facing delivery.