FAQ
This is an exhaustive list of frequently asked questions, including some technical questions.
Basics
What is Jami?
See the Introduction.
What does Jami mean?
The choice of the name Jami was inspired by the Swahili word jamii which means community as a noun and together as an adverb. It was chosen as it reflects the vision for the project: a free/libre program available to all that helps bring communities together, is community supported, and respects the freedom and privacy of the users.
How can I make a bug report?
Please see the Bug report guide.
What makes Jami different from other communication platforms?
Jami does not work like most communication platforms because it is distributed:
Some of the consequences may seem surprising. For instance, since accounts are stored on your device, passwords are optional. However, the most significant practical differences are that you have more freedom and privacy.
What do the green/orange/red status circles next to avatars mean?
On your own account, a red circle means that you aren’t connected to the DHT or offline. You should have a detailled error banner explaining the issue. You may need to check your connection or restart the app.
For contacts, no circle means that no device is detected on the DHT. This contact is unreachable. If an orange circle is present, this means that at least a device is announced on the DHT, so this contact SEEMS to be reachable. You do not have a direct connection towards this peer yet. But they should be able to receive connection requests. Because the device is not directly connected to anything, the device status is unable to be determine for sure. So this can be translated to “Seems to be reachable”, but connection may fail (firewall, deadlock, NAT, etc).
The presence generally have a TTL of 10 min over the DHT and several hours if push notifications are enabled.
If a green circle is present, this means that you’re connected to a device of this peer.
Why is a feature missing on my client?
Not every client implements all features; check the list here to see if your client is missing the feature.
You can make feature requests at https://git.jami.net.
Does Jami support read receipts? Can I turn them on or off?
You can enable or disable read receipts on Android. Other platforms may still be working on this feature. Please see All Features by Client for the current status.
Does Jami support typing notifications? Can I turn them on or off?
Most of the client support sending and receiving typing notifications. You can enable/disable them in the general settings.
Can I make group conference calls?
Yes. You can add Jami contacts to existing calls (audio or video) by clicking the “Add participant” button.
Does Jami have group chats?
Yes, Jami now has ../developer/swarm group chats. However, they are currently experimental and must be manually enabled from application settings. Also, group chats are currently limited to 8 participants (so that bugs and issues could be more easily found and fixed in smaller scenarios, and hopefully lift this limit in the future).
Why aren’t my sent messages showing up on all linked devices?
Before ../developer/swarm, an account’s devices would receive the same messages from contacts if the device was online at the time of the message being sent, but sent messages would not show up on devices other than the one sending the message.
With the introduction of Swarm, conversation histories of new Swarm conversations (including one-on-one conversations) are fully synchronized between all of an account’s linked devices. If you are using an older version of Jami, please upgrade to the latest version with Swarm support. The latest version of Jami is always available from the Download page of the Jami website, at https://jami.net/download/.
To learn more about Swarm, you can read our blog post Synchronizing conversation history with Swarm and see the ../developer/swarm page of the Jami developer manual.
Can I message offline contacts?
With ../developer/swarm conversations, every device stores all a copy of all of the messages in that conversation. If a device (whether your own, or another participant’s) is not available/online when a message is sent, when it comes back online again it will try to fetch any new messages from other online devices/peers and synchronize message history. This can be done whenever at least one other device that has a copy of the new messages is also online.
Note
You can read more about how Swarm conversations are synchronized in our blog post Synchronizing conversation history with Swarm.
If the participants in a conversation are often not online at the same time (for instance, due to timezone differences), one of them might choose to set up Jami on an often-online device that would receive the messages from each participant and relay it to the other(s) whenever they come online. Thus, acting similarly to a “server”, all the while Jami remains distributed by nature.
Where are the configuration files located?
Jami saves its configuration (account, certificates, history) at different locations depending on the platform.
GNU/Linux: global configuration is under
~/.config/jami/dring.yml, and account-specific files are under~/.local/share/jami/. Finally, there is a cache directory at~/.cache/jami/.macOS: the full configuration is under
~/Library/Application Support/Jami/if installed via https://jami.net. The app store version uses~/Library/Containers/com.savoirfairelinux.ring.macos/Data/Library/Application Support/jami/.Android: the full configuration is under
/data/data/cx.ring/(may require root privileges to view or change from outside Jami).Windows: global configuration is under
%AppData%/Local/jami/dring.yml, and account-specific files are under%AppData%/Local/jami/. Finally, there is a cache directory at%USERPROFILE%/.cache/jami/.
Note: audio and video messages are recorded in the local-data in the
folder: sent_data
For files, if a file is saved (right click on the file, then Save) it will be added to the directory you configured in the application settings.
How much bandwidth do I need for calls?
For audio calls, Jami uses about 100 Kbps. For a video call, you need about 2 Mbit/s for medium quality. If your connection is slower, the bitrate will be automatically reduced.
If you are hosting a video conference, you will approximately need an additional 2 Mbps per participant. So, for example for a conference with 10 participants, each participant will need 2 Mbps up & down and the host will need 20 Mbps up and down.
Jami also uses an algorithm to change the consumption depending of the quality of the link. So, the bitrate can have a minimum of 200 Kbit/s and maximum of 6 Mbit/s.
How can Savoir-Faire Linux Inc. (SFL) afford to give Jami away for free? How does SFL make money with Jami?
Savoir-Faire Linux Inc. (SFL) is a consulting company with some R&D projects. Jami is a GPLv3+ project and this will not change. SFL already sells services for several fields of expertise (hosting, developing websites and applications, embedded software, etc). Jami is financed several ways:
Customization contracts
Services for other projects
Visibility
R&D
As a distributed system, Jami incurs very low costs by design
Opt-in collection of anonymized statistics might be added in the future to better understand Jami usage; however, no personal data will be collected.
Account management
What is a Jami account?
A Jami account is an asymmetric encryption key. Your account is identified by a Jami ID, which is a fingerprint of your public key.
What information do I need to provide to create a Jami account?
When you create a new Jami account, you don’t have to provide any private information like an email, address, or phone number.
This is the information you can provide if you choose (it’s all optional):
An avatar.
A display name, which is the name that clients will display for your contact. It can contain special characters.
An optional username, which is a unique identifier that is directly associated with your Jami ID. This username->Jami ID mapping is stored on a server (
ns.jami.netby default, but you can host your own).A password. This password is used to protect the account archive in your device.
More information about Jami accounts is available in the Technical Overview.
Where is my Jami ID?
Your Jami ID should be displayed prominently in whichever app you’re
using. It looks like a long string of letters and numbers; for
example: f2c815f5554bcc22689ce84d45aefdda1bce9146.
Why don’t I have to use a password?
You are not forced to have a password on your account. On a centralized system you would use your password to authenticate with a public server where your account is stored. Someone who knows your password could steal your identity.
With Jami, your account is stored in a folder on your device. The password is only used to encrypt your account in order to protect you from someone who has physical access to your device.
If your device is encrypted, you may not want or need to use a password, and indeed recent versions of Jami don’t ask for an account encryption password by default when creating new accounts.
Note: changing a password will only change the password on the current device and it’s not synced (because their is no server and other devices can be offline anyway).
Why don’t I have to register a username?
The most permanent, secure identifier is your Jami ID, but since these are difficult to use
for some people, you also have the option of registering a username.
Username registration requires a name server, such as Jami’s default
one at ns.jami.net.
If you don’t register a username, you can still choose to register one later at any time.
If you host your own nameserver at example.com, usernames registered
there can be looked up by searching for username@example.com.
Can I change my username?
With the default nameserver (ns.jami.net) you cannot change your
username.
What is the difference between a username and a display name?
You can use your username as an identifier. The username points to your Jami ID, which is your permanent, secure identifier. Two people cannot have the same username.
A display name allows you to choose another name that identifies you to your contacts. Display names can be edited or changed at any time and only your contacts can see them.
How can I back up my account?
There are two ways to back up your account:
Link another device to your account so your account will be on two devices. You can find this option in the account settings page.
Back up the account archive. This file can be found in the account files folder. In some clients, you can export this archive from the account settings.
Can I retrieve my username without my keys?
If you used the default name server at ns.jami.net, you can not.
There is no way to prove it’s your username without your key.
If you use a different name server, there may be a way to move a username to a new Jami ID at the discretion of the administrator of that name server.
For more information about name servers, see ../developer/name-server-protocol.
Can I recover my account if I forget my password?
No. There can not be a traditional account recovery process; you are the only person with access to your data. If you are worried about forgetting your password, please use a password manager.
What happens when I delete my account?
Your account is only stored on your own devices. If you delete your account from each device, the account is gone and you cannot get it back (unless you already made a backup of it earlier). Nobody else can use your account after that.
Your contacts will still have the messages you sent them, but all public record of your account on the DHT will eventually disappear due to absence and lack of activity.
Warning
The default ns.jami.net name server does not
delete any registered usernames – other name servers might (not
recommended), at their administrator’s discretion. So, if you have
an account with a username registered on the default name server
and you delete or lose your account, and did not back up your
account earlier, nobody (including you) will be able to register a
new account with that username again, thus nobody can reach you at
that username anymore.
To avoid losing your account please back it up!
What happens when I link a new device?
When you link a device to your account, your account archive is put on the Jami network for a few minutes. It is protected with a password Jami gives you.
The new device receives your full account certificate with the master RSA keys, but it generates a new device key for signing/encrypting messages.
Advanced
What protocol does Jami use for the end-to-end encryption?
We use TLS 1.3 with a perfect forward secrecy requirement for the negotiated ciphers for calls and file transfers. Messages are encrypted with an RSA key.
What data passes through my machine when I participate in the Jami network?
All these data are encrypted. There is:
ICE descriptors of other Jami users (ICE is a protocol that helps establishing communication between two computers)
certain text messages
accounts currently being linked to a new device, as explained above.
Audio/video streams and some text messages pass through the VOIP protocol. Text messages can be sent either via VOIP or DHT (the distributed network) depending on whether a VOIP communication channel is already open or not.
Why am I able to communicate with myself?
Many users use Jami to transfer data from one machine to another.
Should I enable push notifications?
Push notifications allow Jami to operate in a way more adapted to the context of mobility (energy consumption, data, …). However, for the moment, notifications go through Google’s servers, via the Firebase service. Only one identifier is transferred and it is unusable for anyone who does not have access to your account.
What is a bootstrap server?
A bootstrap server is the entry point of the distributed network.
To enter in a network, Jami must know one other node. This is the
role of the bootstrap. It can be any node in the network, but,
bootstrap nodes are generally always up and available.
The default one in Jami is bootstrap.jami.net.
What is a TURN server? What is STUN?
A TURN server is a relay, and is generally used when two peers can not contact to each other due to some firewall restriction, have NAT without any opened port, and no IPv6.
A STUN server is only used for SIP accounts, and is generally used to get your public IP. For Jami accounts, the DHT already gives this information.
What is DHT proxy?
The DHT proxy is a server that registers on the DHT for you and relays your information to you. Thus, it is the server that will be active on the DHT and will participate in the network, and no longer the target device. Multiple devices can register on the same DHT proxy.
Generally, to transfer data between two peers, there are 3 steps:
Exchange candidates (IPs) via the DHT
Negotiate the best p2p channel between the peers
Transfer data on this socket.
The DHT is only used for the first step.
What if I disable the DHT proxy on Android and what about push notifications?
There is basically 3 modes on how to use the Android application:
With push notifications (DHT proxy must be enabled). This mode supports
notifications for Android (via Google/Firebase, and soon UnifiedPush or Apple/APN). This decrease battery usage, by removing the sync needed with the DHT and without any socket always alive. + Without push notifications but with DHT proxy enabled. This avoids the application synchronizing with other nodes, but “Run in background” MUST be enabled to avoid the operating system killing the application. + Without DHT proxy. In this case, “Run in background” MUST be enabled to avoid the operating system killing the application. The application will synchronize with the other DHT nodes.
I still have issues with the Android application even if battery optimization is disabled
Please read https://dontkillmyapp.com for more details. If it does not solve your issue, you can open a bug report (ideally with a scenario to help us to reproduce and/or logs).
How does the username registration service work?
With the default name server (ns.jami.net), the usernames are
registered on an Ethereum blockchain. If you are a developer, you can
build your own name server with the underlying data storage technology
of your choice (e.g. you could use a SQL database rather than using a
blockchain).
With the default name server, you can look up usernames at
https://ns.jami.net/name/test, where test is a username
for which we are looking for a matching Infohash. Once registered, this name server
does not provide any way to remove the mapping.
Read more about the Jami ../developer/name-server-protocol.
How can I change the timeout for a call?
In the dring.yml file (see Where are the configuration files located?), you can change the ringingTimeout
(in seconds).
How to back up and reimport conversations and accounts
Note
This is only for clients based on LRC (desktop ones).
First you will need to export all your accounts (For GNU/Linux:
Settings => Account => Export account). Then you will
need to copy the database (in ~/.local/share/jami/ for example).
Then on the new device, when you will open Jami for the first time, you have to re-import your accounts via the archive previously saved. This will re-import your settings and contacts (with empty conversations). Then close the client and replace the database with the one previously saved. That’s all!
How secure are you?
We use TLS/SRTP to secure connection and communications over the network.
We implement SRTP over SIP using recommendations described in the following two RFCs:
Typically 2 kinds of sockets are negotiated. One for the control socket, the other for the media sockets.
Typical control session will use the following cipher suite:
(TLS1.3)-(ECDHE-SECP384R1)-(RSA-PSS-RSAE-SHA384)-(AES-256-GCM)
(TLS_ECDHE_RSA_AES_256_GCM_SHA384)
DTLS (fallback) supported:
"SECURE192:-KX-ALL:+ANON-ECDH:+ANON-DH:+SECURE192:-VERS-TLS-ALL:+VERS-DTLS-ALL:-RSA:%SERVER_PRECEDENCE:%SAFE_RENEGOTIATION"
TLS:
"SECURE192:-KX-ALL:+ANON-ECDH:+ANON-DH:+SECURE192:-RSA:-GROUP-FFDHE4096:-GROUP-FFDHE6144:-GROUP-FFDHE8192:+GROUP-X25519:%SERVER_PRECEDENCE:%SAFE_RENEGOTIATION"
Supported crypto suite for the media session are:
AES_CM_128_HMAC_SHA1_80 / SRTP_AES128_CM_HMAC_SHA1_80AES_CM_128_HMAC_SHA1_32 / SRTP_AES128_CM_HMAC_SHA1_32
When do public IPs get exposed?
We can consider three main connectivity scenarios: (1) a classic configuration, (2) behind a VPN, (3) via Tor. As Jami is a p2p application, the reader would probably know that (2) or (3) is a bit mandatory to avoid IP leaking.
Moreover, even if it’s my answer, you can choose to not trust my answer and check the code, or use wireshark or other tools. Generally, I (and as far as I know most other Jami developers) use the first scenario (sometimes the second one), and we surely can not test all the possible networks configurations we would like to, so if you discover a bug, please open an issue.
Anyway, in these 3 scenarios, there are 3 main actions:
sending a message (this will use the DHT);
sending a file (TCP ICE connection as described here: ../developer/file-transfer; and
placing a call (TCP + UDP ICE connection as described here: ../developer/calls).
Classic config
Send a message
The Jami application is running a DHT (https://opendht.net) node on your device. So every operations on the DHT will use your ips. This is why Jami has the option to use a dhtproxy (eg dhtproxy.jami.net), this will avoid to use your node, but will use another node on the network (which will see your ip). Note that your message is not sent directly to the other device. In fact your message is sent on some nodes of the DHT and your contact will retrieve the message on this node. So, your contact don’t see your IP at this step, but the node who get the message will (or they will see the IP of the proxy).
Send a file
As described in the docs, you will send a message with all the IP you know that your peer can contact in an encrypted packet. So, if your peer send you a file or you send a file, your addresses will appear in the ICE message.
Calls
Same as above, the IP is present in the ICE.
Behind a VPN
Send a message
The IP of your VPN will be used by the DHT node. If you want a proof,
you can compile dhtnode and run the la command to get your public
detected address. This is what I got:
./tools/dhtnode -b bootstrap.jami.net
Bootstrap: bootstrap.jami.net:4222
OpenDHT node be58fdc9f782269bfc0bbfc21a60bca5f02cb881 running on port 54299
(type 'h' or 'help' for a list of possible commands)
>> la
Reported public addresses:
IPs OF MY VPN
So, if you don’t use a proxy, your VPN addresses will be used for using the DHT. If you use a dhtproxy, the dhtproxy will see your VPN addresses
Send a file
Same as above, the ICE will contains: + addresses from your LAN + public address of your VPN + TURN address if TURN is enabled
Do a call
Same as above, your public address is replaced by your VPN address. You can see it in the logs from daemon. See Logs.
Tor
Send a message
Tor basically does not supports UDP. This means that you can not use your DHT node locally, you MUST use a DHTProxy. That proxy will see the Exit node.
Send a file
I prefer a proof that any description. So, I did a file transfer with Jami + TOR. This is what I see in the logs for the remote:
[1574218330.556|10688|p2p.cpp :241 ] [Account:93a03f519f394143] add remote ICE candidate: Hc0a8c801 1 TCP 2130706431 192.168.200.1 33293 typ host tcptype passive
[1574218330.556|10688|p2p.cpp :241 ] [Account:93a03f519f394143] add remote ICE candidate: Hc0a8c801 1 TCP 2130706431 192.168.200.1 9 typ host tcptype active
[1574218330.556|10688|p2p.cpp :241 ] [Account:93a03f519f394143] add remote ICE candidate: Hc0a80103 1 TCP 2130706431 192.168.1.3 33293 typ host tcptype passive
[1574218330.556|10688|p2p.cpp :241 ] [Account:93a03f519f394143] add remote ICE candidate: Hc0a80103 1 TCP 2130706431 192.168.1.3 9 typ host tcptype active
[1574218330.556|10688|p2p.cpp :241 ] [Account:93a03f519f394143] add remote ICE candidate: R33fe279d 1 TCP 16777215 51.254.39.157 27427 typ relay tcptype passive
[1574218330.556|10688|p2p.cpp :241 ] [Account:93a03f519f394143] add remote ICE candidate: Sc0a8c801 1 TCP 1694498815 185.220.101.24 33293 typ srflx tcptype passive
The first ones are some 192.168.x.x so we don’t care. 51.254.39.157 is the TURN address in France (my device is in the Canada). 185.220.101.24 is the Tor exit node:
inetnum: 185.220.101.0 - 185.220.101.127
netname: MK-TOR-EXIT
Do a call
This will not work (actually, you can create the SIP control connection because it’s a TCP connection), but medias are negotiated in UDP, so this will fail.
What ports does Jami use?
Jami works as a server and gets new ports for each connections (randomly bound). These are the ranges that can be used for each component:
dht: UDP [4000, 8888]
audio: UDP [16384-32766]
video: UDP [49152-65534]
SIP Control: UDP/TCP randomly bound
Note: if UDP is blocked, a dhtproxy can be used to use TCP instead. Note that medias will not work cause it only supports UDP.
So for ufw, we recommend running sudo ufw default allow outgoing.
For now, you can not specify a specific range to configure ports used by Jami. The inbound traffic can be controlled without issue, Jami should work and can use a TURN server if needed.
If you run your own proxy or nameserver:
dhtproxy, nameserver: TCP [80-100], 443
If you run your own TURN server:
TURN/STUN: TCP+UDP 3478, 5349
Can I use Jami in a local network (LAN) without internet access?
Yes! Thanks to Jami’s architecture, Jami users on a local/private network can communicate among themselves using Jami, without requiring any outside connectivity such as the internet.
To do so, from Jami’s Account settings open Advanced account
settings. There, enable the Enable local peer discovery
setting. Additionally, you may want to manually set the bootstrap
node’s address (default: bootstrap.jami.net) to the IP address of
another device on your network that also runs Jami and/or an OpenDHT
node.
Note
If you will use this Jami account for communicating only with only with other devices on the same local/private network, you can disable TURN if you wish. If you do so, and later you decide to use this account also for communicating with other Jami devices outside your network, don’t forget to enable TURN again, as it helps Jami work around issues with some overly restrictive firewalls.
How can I configure the codecs even more?
Codecs can be configured via a file. In the configurations files, you
can create a file called encoder.json like this:
{
"libx264": {
"profile": 100,
"level": 42,
"crf": 20,
"preset": "ultrafast"
},
"h264_vaapi": {
"low_power": 1
},
"libopus": {
"application": "voip"
}
}
or:
{
"libopus": {
"bit_rate": 128000
}
}
This file is located in the same directory as dring.yml.
To check which options are supported, use the command ffmpeg -h
encoder=[encoder_name], where encoder_name can be any of
libx264, libvpx, mpeg4, h263, libopus,
libspeex, g722, pcm_alaw, or pcm_mulaw (the FFmpeg
names for all of Jami’s supported encoders).
How can I configure the audio processor?
An audio processor allows Jami to clean up and process your microphone’s audio.
It can remove echo, reduce noise, and equalize your microphone’s volume.
Additionally, it can detect when you’re speaking and send this information to participants in your call.
The audio processor settings can be set in your dring.yml file.
See this section to find where this file is located.
The relevant preference keys are:
audioProcessor, which configures which audio processor to use. The valid options are:webrtc: the WebRTC Audio Processing libraryspeex: the Speex DSP librarynull: disables audio processing (though your system echo canceller may still be used, see below)
echoCancel, which configures how echo cancelling should be done. The valid options are:auto: try to use your operating system’s echo canceller (if it exists), otherwise fall back to the chosen audio processor’s echo cancelleraudioProcessor: only use the chosen audio processor’s echo cancellersystem: only use your operating system’s echo cancellernull: don’t do any echo cancelling
noiseReduce,true/falseto set noise reduction on the audio processorautomaticGainControl,true/falseto set automatic gain control on the audio processorvoiceActivityDetection,true/falseto set voice activity detection on the audio processor